(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Biodiversity Heritage Library | Children's Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Flight evaluation of an advanced technology light twin-engine airplane (ATLIT)"

NASA CONTRACTOR 

J; 

f REPORT 



esi 

CO 

bo 



um 

^-«« 








LO^N copy: return to 

AFWL TECHN5CAL LIBRARY 
KIRTLAND AFB, N. M. / 



SLIGHT EVALUATION OF 
|.N ADVANCED TECHNOLOGY 
LIGHT TWIN-ENGINE AIRPLANE (ATLIT) 

I 

^ruce J, Holmes 



Prepared by 

InIVERSITY of KANSAS 

I 

awrence, Kaos. 66045 

41' 

Ir L.angley Research Center 

I 

tATIONAL AERONAUTICS AND SPACE ADMINISTRATION 



N '-7 • 



WASHINGTON, D. C. • JULY 1977 



i 



TECH LIBRARY KAFB. NM 

lillNiililll 



ODtivia 



2. Government Accession No. 



LI. Report No. 
NASA CR-2832 
4. Title and Subtitle 

FLIGHT EVALUATION OF AN ADVANCED TECHNOLOGY LIGHT TWIN-ENGINE 
AIRPLANE '(ATLIT) 



7. Author(s) 

Bruce J. Holmes 



9. Performing Organization Name and Address 

The University of Kansas 
Lawrence, Kansas 66045 



12. Sponsoring Agency Name and Address 

National Aeronautics and Space Administration 
Washington, DC 20546 



3. Recipient's Catalog No. 



5. Report Dati 



SSfy'S^T 



6. Performing Organization Code 



8. Performing Organization Report No. 



10. Work Unit No. 
505-10-11-03 



11. Contract or Grant No. 
NGR 17-002-072 



13. Type of Report and Period Covered 

Contractor Report 



14. Sponsoring Agency Code 



15. Supplementary Notes 
Topical report . 
Langley Technical Monitor : 



Harold L. Crane 



16. Abstract 

The project organization and execution, the airplane description and performance predictions, 
and the results of the flight evaluation of an advanced technology light twin-engine airplane 
(ATLIT) are presented. The ATLIT is a Piper PA- 34-200 Seneca I modified by the installation of 
new wings incorporating the GA{W)-1 (Whitcomb) airfoil, reduced wing area, roll-control spoilers, 
and full-span Fowler flaps. The conclusions for the ATLIT evaluation are based on complete stall 
and roll flight-test results and partial performance test results. Stalling and the rolling 
characteristics met design expectations. The cruise and climb performance did not meet the 
predictions. Climb performance was penalized by extensive flow separation in the region of the 
wing-body juncture. Cruise performance was found to be penalized by a large value of zero-lift 
drag. Calculations showed that, with proper attention to construction details, the improvements 
in span efficiency and zero-lift drag would permit the realization of the predicted increases in 
cruising and maximum rate-of-cl imb performance. Plans for future tests include the following 
topics: climb performance will be documented with the flow separation at the wing-body juncture 
cleaned up; the noise and performance characteristics of a set of supercritical propellers will be 
measured; testing is planned for ATLIT in the NASA-LaRC full-scale (30- by 60-foot) wind tunnel 
in the fall of 1976; and, after the wind-tunnel tests, ATLIT will return to flight status for 
evaluation of the final wind-tunnel optimized configuration 



17. Key Words (Suggested by Author{s)) 

Advanced technology light twiii; 6A(W)-1 airfoil; 
LS-1 0417 airfoil; spoilers; full -span Fowler 
flaps; airspeed calibration; aerodynamic 
parameter extraction; airplane performance 
measurements 



18. Distribution Statement 



Unclassified 



Unlimited 

Subject Category 02 



19. Security Qassif. (of this report) 

Unclassified 



20. Security Classif. (of this page) 

Unclassified 



21. No. of Pages 

287 



22. Price* 



$9.25 



For sale by the National Technical Information Service, Sprlnsfield, Virginia 22161 



LIST OF ACRONYMS 

ACD: Analysis and Computation Division (at LaRC) 

ADTRAN: Analog to Digital Translation 

ATLIT: Advanced Technology Light Twin 

eg: Center of Gravity 

CRINC: (KU) Center for Research, Inc. 

DOC: Direct Operating Cost 

FM: Frequency Modulated 

FRD: Flight Research Division (at LaRC) 

FRL: Flight Research Laboratory (at KU) 

GA(W): General Aviation (Whit comb) (airfoil) 

ILS: Instrument Landing System 

KU: University of Kansas 

LaRC: (NASA)-Langley Research Center 

LS-1: Low Speed (family of airfoils) 

mac: Mean Aerodynamic chord 

NACA: National Advisory Committee on Aeronautics (became NASA in 1957) 

NASA: National Aeronautics and Space Administration 

NCSU: North Carolina State University 

PAM: Pulse Amplitude Modulated 

QA: Quality Assurance Office (LaRC) 

RMS: Root Mean Square 



111 



SOPB: Safety and Operating Problems Branch (in FRD) 

V/STOL: Vertical /Short Takeoff and Landing 

W.S. : Wing Station 

WSU: Wichita State University 



IV 



TABLE OF CONTENTS 

Pa^e 

LIST OF ACRONYMS iii 

TABLE OF CONTENTS v 

LIST OF FIGURES viii 

LIST OF TABLES xiv 

LIST OF SYMBOLS xvi 

CHAPTER 

1 INTRODUCTION . . 1 

2 PROJECT CHRONOLOGY AND MANAGEMENT 7 

2.1 Project History 7 

2.2 Project Support Organization 12 

2.3 Project Budget 18 

2.4 Project Schedule . 25 

3 AIRPLANE MODIFICATIONS 28 

3.1 General Seneca/ATLIT Description and Comparison , . 28 

3.2 ATLIT Design Description 29 

3.2.1 ATLIT Planform Changes 35 

3.2.2 Roll -Control Spoilers 39 

3.2.3 Full -Span Fowler Flaps 49 

3.2.4 GA(W)-1 Airfoil 52 

3.3 Supercritical Propellers 58 



Page 

FLIGHT-TEST PROGRAM 62 

4.1 Flight-Test Program Objectives and Planning ... 52 

4.2 Flight Test Instrumentation , . 54 

4.2.1 ATLIT Instrumentation Recording Package . . 54 

4.2.2 Nose Boom Installation 55 

4.3 ATLIT Flight Envelope 72 

FLIGHT-TEST RESULTS 76 

5.1 Position Error Calibrations for Static Pressure and yg 
Angle of Attack 

5.1.1 Static Pressure Calibrations 80 

5.1.2 Angle of Attack Calibrations so 

5.2 Stall Speeds and Characteristics 83 

5.2.1 Predictions 83 

5.2.2 Methods and Data Reduction 86 

5.2.3 Results 88 

5.3 Spoiler Rolling Characteristics 92 

5.3.1 Spoiler System Development 92 

5.3.2 Methods and Data Reduction 94 

5.3.3 Results 97 

5.4 Cruise and Single/Multi-Engine Climb Performance . 121 
5.4.1 Predictions 121 

5.4.1.1 Method A: Performance Predictions. 121 

5.4.1.2 Method B: Lift, Drag, and 

Performance Predictions .... 130 

5.4.1.3 Method C: Lift, Drag, and 

Performance Predictions .... 140 



VI 



Page 

5.4.2 Methods, Data Reduction, and Results ... 145 

5.5 Pilots Descriptions of Stability and Handling 

Qualities 157 

5.5.1 Pilot A Comments on ATLIT Flying Qualities . 157 

5.5.2 Pilot B Comments on ATLIT Flying Qualities . 165 
6 CONCLUSIONS AND RECOMMENDATIONS 179 

6.1 Conclusions 179 

6.2 Recomnendations 181 

6.2.1 Supercritical Propeller Evaluation .... 182 

6.2.2 Full-Scale Wind-Tunnel Plans 182 

6-2.3 Final Flight Evaluation 184 

REFERENCES 185 

APPENDIX A: Performance Parameter Extraction Method with Error 

Analysis 190 

APPENDIX B: Position-Error Calibrations for Static Pressure 

Systeln and Angle-of -Attack Vane 229 

APPENDIX C: Prediction of Roll Damping Derivatives 257 



vn 



LIST OF FIGURES 

Paae 

The Advanced Technology Light Twin-Engine 

Airplane (ATLIT) 3 

Three-View of Redhawk Compared to Original Cessna 

Model 177 Cardinal 8 

Milestones in General Aviation NASA Grant Research 

At the University of Kansas 13 

ATLIT Project Support Organization 14 

Data Reduction Flow Diagram 19 

ATLIT Flight-Test Program Timetable 26 

Planview Comparison of the ATLIT and the Unmodified 

Piper PA-34-200 Seneca I 30 

ATLIT Three-View 32 

Spoiler Installation Detail 40 

Comparison of Predicted and Wind-Tunnel -Measured 
(Reference 17) Spoiler Hinge Moments for ATLIT .... 

(a) Fowler Flap Nested ^3 

(b) Fowler Flap Deflected 40° 44 

3.5 Relationship Between Spoiler Deflection and 

Control -Wheel Position 46 

3.6 The Effect of Opening (Positive) Hinge Moments 

on Wheel Forces 47 

3.7 ATLIT Spoiler Leakage Path With Seals 50 

3.8 ATLIT airfoil and flap geometry 53 

3.9 Comparison of the GA(W)-1 and 652-415 Airfoil Shapes . . 55 



Fi! 


fjure 


1 


.1 


2, 


.1 


2, 


.2 


2. 


.3 


2. 


.4 


2. 


.5 


3. 


.1 


3. 


2 


3. 


3 


3, 


4 



vm 



Figure Page 

3.10 Comparison of 6A(W)-1 and 652-415 Section 

Characteristics 55 

3.U ATLIT Wing Templates 59 

3.12 Supercritical Propeller Planform and Cross-Section. . . eV 

4.1 ATLIT Noseboom Detail 67 

4.2 Time Dependent Characteristics of the ATLIT Pi tot 

Pressure Measuring System 70 

4.3 Time Dependent Characteristics of the ATLIT Static 

Pressure Measuring System 71 

4.4 Comparison of PA-34 and ATLIT Planforms 73 

4.5 Airplane CG Envelope for ATLIT and PA-34 Mac's ... . 74 

5.1 Devices for Wing/Body Flow Attachment 77 

5.2 Static Pressure Position-Error Calibrations 81 

5.3 Comparisons of Fowler Flap Effectiveness from Wind- 

Tunnel and Flight Tests 84 

5.4 ATLIT Stall Time History (Flaps Up, Approach Power) . . 87 

5.5 Variation of ATLIT Spoiler System Stretch with 

Airspeed (Flaps Up, Spoiler Leakpath Sealed) .... 98 

5.6 Roll Response Time Histories (6^ = 30°, 78 kt IAS) . . 

(a) Large Spoiler Deflection lOO 

(b) Small Spoiler Deflection 101 

5.7 Rolling Velocity as a Function of Spoiler Deflection . . 

(a) <5^ = 0° 103 

(b) 6^ = 10° 104 

(c) 6^ = 20° 105 

(d) 6f = 30° 106 

(e) 6f = 37° 107 



IX 



Figure Page 

5.8 Roll Helix Angles for Varying Spoiler Inputs .... 

(a) 6f = 0° 109 



(b) 5f = 10° 110 

(c) 6f = 20° Ill 

(d) 6. = 30° 112 



(e) 6^ = 37° 113 

5.9 Roll Helix Angles for Varying Airspeed 

(6^ = 0°. 10°, 20°. 30°, 37°) II4 

5.10 Variation of Rolling Moment Coefficient with Airspeed 

(a) 6^ = 0° • 116 

(b) 6^ = 10° 117 

(c) 6f = 30° 118 

5.11 Pilot Ratings for Nonlinear Roll Response Variations 
(Reference 19) 12o 

5.12 Predicted ATLIT Drag Polar and Wind-Tunnel Measured 
Light-Twin Drag Polar (Reference 36) I37 

5.13 Assumed Power Available for ATLIT and Seneca 

Performance Predictions I39 

5.14 Comparison of Measured and Predicted Lift 

Characteristics for the ATLIT and the Seneca .... 141 

5.15 Predicted Drag Polars for the ATLIT and the Seneca 
(Ideal Skin Friction, 2-D Profile Drag Data Used 

with Standard Roughness) 142 

5.16 Comparison of Predicted Power Required for the 

ATLIT and the Seneca I43 

5.17 Variation of Shaft Power Required with Airspeed 

for ATLIT (6^ = 0°, 10°, 30°) 146 



Figure Page 

5. 18 ATLIT Lift Characteristics : (Power l^or Level F] ight; 

6^ = 0°. 10®. 3a°) 148 

5.19 Power On, Trlnmed Drag Polars for ATLIT 

(6^ = 0*^, 10°, 30°) • 151 

5.20 Linearized, Power-On Drag Polar for ATLIT 

(6f = 0°. 10°, 30°) 152 

3/2 

5.21 Variation of (C|_ /C^) with Lift Coefficient 

for ATLIT (6^ = 0°, 10°, 30°) 154 

5.22 Variation of Rate of Climb with Airspeed; 
Single-Engine and Multi-Engine {6^ = 0°, 5°, 10°) 

for ATLIT 155 

5.23 ATLIT Stall Time Histories 173 

A.l The Effect of Weight Biasing 219 

(a) Extracted Power Coefficient 

(b) Extracted Drag Coefficient 

A. 2 The Effect of Altitude Biasing 220 

(a) Extracted Power Coefficient 

(b) Extracted Drag Coefficient 

A. 3 The Effect of Airspeed Biasing 221 

(a) Extracted Power Coefficient 

(b) Extracted Drag Coefficient 

A. 4 The Effect of Airspeed Biasing 222 

(a) Extracted Power Coefficient 

(b) Extracted Drag Coefficient 



XI 



Figure 
A. 5 



A. 6 



A. 7 



A. 8 



A. 9 



A. 10 



B.l 
B.2 

B.3 
B.4 



The Effect of Acceleration Biasing 



(a 
(b 



The Effect of Acceleration Biasing 



(a 
(b 



(a 

(b 

Th 

(a 

(b 

The 

(a 

(b 

The 

(a 

(b 



Extracted Power Coefficient 
Extracted Drag Coefficient 



Extracted Power Coefficient 
Extracted Drag Coefficient . . 
The Effect of Pitch Angle Biasing - 
Extracted Power Coefficient . 
Extracted Drag Coefficient . 
Effect of Pitch Angle Biasing . 
Extracted Power Coefficient . . 
Extracted Drag Coefficient . , 
Effect of Angle of Attack Biasing 
Extracted Power Coefficient . 
Extracted Drag Coefficient . 
Effect of Angle of Attack Biasing 
Extracted Power Coefficient . 
Extracted Drag Coefficient . 



ATLIT Trailing Anemometer Installation Detail 

Induced Velocity Contours (from Reference B.4) 
and Trailing Anemometer Locations .... 



Wallops Tower Flyby Airspeed Calibration Dimensions 

Effect of Altitude Error on Tower Flyby 

Airspeed Calibration 



Page 
223 



224 



225 



226 



227 



228 



246 

247 
248 

249 



xn 



Figure Page 

B.5 Comparison of Tower Flyby Anemometer (Static and 
Continuous Runs) Static Pressure Position Error 

Calibrations (6^ = 0*^) 250 

B.6 Effect of Power on Static Pressure Position Error 

Calibrations (Trailing Anemometer Continuous Runs) . . 250 

B.7 Effect of Flap Deflection on Static Pressure Position 
Error Calibration (Trailing Anemometer Continuous Runs) 
Power On 251 

B.8 Flight-Test Lift Data for True (Geometric) Angles of 
Attack (6^ = 0° (Spoiler Leak Path Sealed), 10°, 30°) 
CG ^ 15% mac 252 

B.9 Flight-Test Lift Data for Indicated Angles of Attack 
(6^ = 0° (Spoiler Leak Path Sealed), 10°, 30°) 
CG - 15% mac 253 

B.IO Angle-of-Attack Position Error Calibrations 

(a) Flaps Up 254 

(b) Flaps 10° 255 

(c) Flaps 30° 2156 

C.l Conditions and Assumptions for Estimating Engine Nacelle 

Contribution to Airplane Roll Damping Derivatives ... 263 

C.2 Variation of Two- Dimensional Lift Curve Slope with 

Angle of Attack (6^ = 0°, 10°, 20°, 30°, 40°) .... 264 

C.3 Predicted Airplane Roll Damping Derivatives 

(6f = 0°, 10°, 30°) 265 



xm 



L 


,1 


2. 


,1 


2. 


.2 


2. 


.3 



LIST OF TABLES 



Table Page 

Industry Utilization of Advanced General Aviation 

Technology 5 

Comparison of Redhawk and Cessna Cardinal Design 

Specifications (From Reference 3) • . 10 

Total ATLIT Project Budget 20 

Total Flight Program Operating Costs for 85 Hours of 

Research Flying 24 

3.1 Comparision of ATLIT and PA-34-200 Seneca I Design 
Specifications 31 

3.2 ATLIT Engineering Design Drawings 33 

3.3 Fowler Flap Slot Dimensions for Wind Tunnel and 

Flight Testing 51 

4.1 ATLIT Instrumentation Parameters and Accuracies ... g^ 

4.2 Effects of Flow Angularity on the Pitot-Static 

Measurements (From Reference 28, M = 0.6) gg 

5.1 Angle-of -Attack Calibration Equations 32 

5.2 Comparison of ATLIT and Seneca Stall Speeds and 

Maximum Trimmed Lift Coefficients 39 

5.3 Configuration and Airspeeds for Spoiler Roll Tests . . gg 



5.4 Comparison of Piper Seneca I Performance with 
Predictions for ATLIT 



122 



5.5 Comparisons of Predicted Performance Parameters 

for the Standard Seneca and ATLIT (W = 1.87 kN, er = 1). 127 

5.6 ATLIT Drag Buildup 135 

5.7 Handling Qualities Rating Scale 158 



XIV 



Table Page 

B.l Summary of Static-Pressure Position-Error 

Calibration Method Accuracies 244 

B.2 Configuration/Airspeed Combinations for ATLIT 

Static Pressure- System Calibration Tests 245 

C.l Wing Geometry with Flaps Deflected 262 



XV 



LIST OF SYMBOLS 
Symbol Definition Dimension 

2 
A Aspect ratio, b_ 

S 

a Acceleration along airplane g 

longitudinal axis 

b Wing span m (ft) 

c Wing chord m (ft) 

C Constant 

c Specific fuel consumption N / lb 

Watt-hr \HP-hr^ 

c. Two-dimensional zero-lift drag 
coefficient 

Cq Three-dimensional zero-lift drag 
coefficient 



Cm Hinge moment coefficient H 



'H 



a 



'£ 



q.S.c 



c.„ Two-dimensional lift coefficient 
C|_ Three-dimensional lift coefficient 



a 8C 



Two-dimensional lift-curve slope, rad" (deg' ) 



i 



3a 



C -1 -1 

L Three dimensional lift curve slope, rad (deg ) 

ac, 



3a 



C^,"* Airplane rolling moment coefficient, _L 



q^ 



XVI 



I 



S^onbol Definition Dimension 

Cn Variation of rolling moment coefficient rad" (deg" ) 

^s ^^i 

with spoiler deflection, tt— 

C^ Airplane roll damping, —7^177 ^^^ (^sg ) 

P 



<ij 



c Two-dimensional zero-lift pitching 

moment coefficient 



C Wing root chord m (ft) 

C. Wing tip chord m (ft) 

d Lateral distance between fuselage m (ft) 
centerline and nacelle center! ine 

D Drag N (lb) 

e Airplane (Oswald's) efficiency factor 

2 2 
g Acceleration of gravity m/sec (ft/sec ) 

h Altitude m (ft) 

Ah 

c Spoiler projection height 

H Hinge moment N-m (ft-lb) 

k Constant 

k Distance behind airplane in terms of 
wing spans (k = x/b) 

k^ Distance below airplane in terms of wing 
spans (k^ = z/b) 

L Lift N (Ibf) 

L" Airplane rolling moment ^~^ (ft-lbf) 



xvn 



Sjmibol 

it/d 
L/D 



Definition 

Airplane horizontal tail length 
(elevator hinge line to airplane 
center of gravity) 

Two-dimensional lift to drag ratio 

Three-dimensional lift to drag ratio 



Dimension 



m (ft) 



M 

mac 
n 

n 



P 
P 

pb/2V 

q 

Q 

RN 
R 

R 
S 



Mach number, n— 
^a 

Mean aerodynamic chord 
Load factor 

3n 
Gust load factor, r— 

Roll rate, ^ 

Static pressure 

Total pressure 

Airplane power 

Coefficients in polynomial expression 
of power available 

Roll helix angle 

2 
Dynamic pressure, 1/2 pV 

Pitch rate, ^ 

Reynolds number 

Gas constant for air 

Yaw rate, j±- 

Reference wing area 
Nacelle planform area 



m (ft) 

g/rad 

deg/sec 

pa (psf) 
pa (psf) 
watts (HP) 

rad 

pa (psf) 

deg/sec 



deg/sec 

\/ (ft^) 
m^ (ft^) 



xvm 



Symbol 

S 

T 

t 

T 

t 
c 



^i 



'NE 



x/D 



Definition 
Error term 
Temperature 
Time 
Thrust 

Wing thickness ratio 

Component of velocity along airplane 
longitudinal axis 

True airspeed 

Local speed of sound 

Calibrated airspeed 

Indicated airspeed (= V "*) 

Never exceed airspeed 

Airplane power-off stall speed in 
landing configuration 

Airplane weight 

Component of velocity along airplane 
vertical axis 



Dimension 

°K. °C. (°R, °F) 

sec 

N (lb) 

% 

m/sec (ft/sec) 

knots (mph) 
m/sec Cft/sec) 

knots (mph) 

knots (mph) 

knots (mph) 

knots (mph) 

N (lb) 

m/sec (ft/sec) 



Distance between accelerometer location m (ft) 
and airplane center of gravity (longitudinal) 

Ratio of distance along airplane 
longitudinal axis to maximum fuselage 
diameter 



Distance between accelerometer location 
and airplane center of gravity (vertical) 



GREEK SYMBOLS 



True (geometric) angle of attack 
(referenced from fuselage centerline) 

Measured (indicated) angle of attack 
(referenced from fusel aae centerline) 



m (ft) 

deg 
deg 



XIX 



S:ymbol 


Definition 


Dimension 


3 


Sideslip angle 


deg 


3' 


Measured sideslip angle 


deg 


A 


Difference or increment 




6 


Deflection 


deg 


6* 


Boundary layer displacement thickness 


cm (in) 


r 


Circulation -^ 


2 2 
mvsec (ft /sec) 


y 


Flight path angle 


deg 


Y 


Ratio of specific heats for air 




X 


^t 
Taper ratio, 7^ 

^r 




n 


Efficiency 




* 


Bank angle 


deg 


il' 


Yaw angle 


deg 


p 


Air density 


N/m^ (slugs/ft^) 


e 


Pitch attitude 


deg 



Airplane pitch attitude 

Density ratio, -^ 
^0 



deg 



T 


Time constant 


SUBSCRIPTS 




a 


Airplane 


A 


ATLIT 


A 


Airplane 


ac 


Aerodynamic center 


b 


Barograph 


c 


Calibrated 



sec 



XX 



Symbol 


Definition 


c 


Corrected 


c/4 


Quarter chord 


f 


Flap 


fus 


Fuselage 


g 


Gross 


i 


Indicated 


in 


Installed 


ind 


Indicated 


m 


Measured 


max 


Maximum 


min 


Minimum 


n 


Nacelle 


s 


Spoiler 





Zero lift value 





Sea level value 





Starting or reference value 


P 


Propeller 


ref 


Reference 


s 


Stagnation 


s 


Static 


S 


Seneca 


s 


Spoiler 


t 


Trim 


t 


Total 



XXI 



Symbol 


Definition 


t 


(Horizontal) tail 


te 


Trailing edge 


te,o 


Trailing edge at zero wing incidence 


% 


(Vertical) tail 


w 


(Control) wheel 


w 


Wing 


SUPERSCRIPTS 



* f\\l 

Dotted symbols represent time derivatives (e.g., V = -g^) 

Primed symbols represent measured or indicated values, 
with position error 



xxn 



CHAPTER 1 

INTRODUCTION 

This report presents the project organization and execution, the 
airplane design description, the airplane performance predictions, and 
the results of the flight evaluation of an advanced technology light 
twin-engine airplane (ATLIT). The results cover the period from the ATLIT 
first flight in October 1974, to June 1976. Some pre-ATLIT historical 
notes are also included. 

The flight-test results include stall characteristics, spoiler 
roll performance, cruise and single/multi-engine climb performance, and 
pilot comments on stability and handling qualities. Planned tests which 
are not in the scope of this report include takeoff and landing performance 
evaluation, stability derivative determination, supercritical propeller 
evaluation, and full-scale (30- by 60-foot) wind-tunnel tests. 

The ATLIT is the second airplane designed and constructed as part 
of a general aviation research program at the University of Kansas (KU) 
Flight Research Laboratory (FRL), sponsored by grants (NGR 17-002-072) 
from the National Aeronautics and Space Administration (NASA), Langley 
Research Center (LaRC). The airplane which preceded ATLIT in development 
Is the Redhawk, a modified Cessna Cardinal (references 1, 2, and 3). The 
object of the research under these grants has been to apply existing 
jet-transport wing technology and advanced airfoil technology to general 
aviation airplanes for the purpose of improving safety, efficiency, and 
utility. 



The ATLIT is a Piper PA- 34-200 Seneca I with the following 
modifications: 

1. Wing planform modified for cruise efficiency with 
taper, reduced area, and increased aspect ratio. 

2. Full -span Fowler trailing edge flaps. 

3. Spoilers for roll control. 

4. 6A(W)-1, general aviation (Whitcomb) 17-percent thick 
airfoil . 

5. Ground-adjustable wing incidence. 

6. Advanced-technology propellers incorporating a 
supercritical airfoil. 

The airplane appears in figure 1.1. 

The ATLIT project is a multi-purpose program. Performance improvements 
throughout the flight envelope are sought, with emphasis on the enhancement 
of the safety of light, twin-engine airplanes by increasing the single-engine 
climb performance through aerodynamic changes. Preliminary design estimates 
(reference 4 and unpublished data ) indicate that the airplane modifications 
mentioned above would result in Improvements to both the single-engine rate 
of climb and the cruise performance. The ATLIT wing was designed to take 
advantage of the lov( profile drag characteristics (at climb conditions) 
of the GA(W)-.l airfoil (reference 5), and of the lower induced and profile 
drag characteristics (at both climb and cruise conditions) of the modified 
wing planform. The cruise-optimized planform logically led to the use of 
full -span Fowler flaps for acceptable landing speeds with roll control 



1. Conceptual Design of an Advanced Technology Light Twin Aircraft, 
Phase I Report: Prepared by Robertson Aircraft Corporation, and 
the University of Kansas Center for Research, Inc., for NASA 
Langley Research Center under NASA Grant NGR 17-002-072, 1972. 




Figure 1.1.- The advanced technology light twin-engine airplane (ATLIT) 



provided by spoilers. Along with predicted performance improvements, the 
wing modifications would result in improved ride quality (due to higher 
wing loading) and would permit lighter structural wing weight (due to 
thicker wing sections and reduced wing area). 

Stimulated by the ATLIT project, and in fulfilling its role to 
provide technology to aircraft manufacturers, NASA has undertaken the 
development of a new family of low-speed airfoils for use on general 
aviation aircraft. This new airfoil family is a product of the 
development of computational methods for designing optimized airfoil 
shapes (reference 6). Application of the GA(W)-1 section to the ATLIT 
wing represents the first effort to verify the characteristics of a 
computer-designed airfoil in flight. This flight verification closes 
the loop in the computer/wind-tunnel /flight hardware design process. In 
addition to documenting the new airfoil characteristics, ATLIT provides 
data on the use of full -span Fowler flaps combined with roll -control 
spoilers on the GA(W)-1 wing. Although the performance characteristics 
of these roll -control and high-lift devices have been documented in the 
literature, there currently exist: little practical experience and data 
(spoiler hinge-moments and flap effectiveness with the new GA(W)-1 
airfoils, for instance,) concerning their application to modern general 
aviation aircraft. The complete documentation of the ATLIT airplane 
characteristics will make such information available to the U.S. industry. 

The interest shown by the general aviation industry in the 
aerodynamic devices used on the two KU-modified airplanes can be 
illustrated by table 1.1. The ten airplanes listed in the table have 
flown with or are being designed utilizing some combination of the. devices 
discussed. 



^==^^^ 



TABLE 1.1.- INDUSTRY UTILIZATION OF ADVANCED GENERAL AVIATION TECHNOLOGY 





GA(W) 


ROLL CONTROL 


FOWLER 


AIRPLANE 


AIRFOIL 


SPOILERS 


FLAPS 


BEECHCRAFTPD-285 


X 






ROBERTSON/SENECA 




X 


X 


BEDE 5 (JET) 


X 






BEDE 5 (PROP) 


X 






AMERICAN JET "HUSTLER" 


X 


X 


X 


ROBERTSON/CESSNA 400-SERIES 






X 


ROBERTSON/BONANZA 




X 


X 


PIPER TRAINER 


X 






RUTAN/VARI-EZE 


X 






CESSNA 441 






X 



The ATLIT flight-test program is being conducted at NASA Langley 
Research Center, in Hampton, Virginia. Many individuals support the 
project directly or indirectly. Those directly contributing to the 
preparation of this project report are acknowledged below: 

Mr. Harold L. Crane, NASA LaRC (Project Technical Monitor): 
Flight data analysis on spoiler roll characteristics (Chapter 5.3). 

Mr. Joseph H. Judd, NASA LaRC: Flight data analysis on 
cruise and single/multi-engine climb performance (Chapter 5.4). 

Mr. Robert A. Champine and Mr. Philip W. Brown, NASA LaRC 
(research pilots): Pilot comments on ATLIT stability and 
handling qualities (Chapter 5.5). 

Mr. Robert T. Taylor, NASA LaRC: Performance predictions 
(Chapter 5.4). 

Mr. Laurence K. Loftin, Jr., NASA LaRC: Performance 
predictions (Chapter 5.4). 

Dr. Frederick 0. Smetana, North Carolina State University: 
Lift, drag, performance, and stability predictions (Chapter 5.4), 
and drag/power parameter extraction method (Appendix A). 

Mr. Bradley J. Vincent, Embry-Riddle Aeronautical University: 
Roll damping derivative predictions (Appendix C). 

Flight testing of ATLIT will continue from the date of this report 
until early fall 1976, when the airplane will enter the full-scale 
(30- by 60- foot) wind tunnel at LaRC. Flight- test results for this period 
will be presented in NASA and technical society publications. 

Commercial products and/or names of manufacturers are used in this 
report documenting the flight evaluation results of ATLIT. These commer- 
cial products and/or names of manufacturers do not constitute official 
endorsement, expressed or implied, of such products or manufacturers by 
the National Aeronautics and Space Administration. 



CHAPTER 2 

PROJECT CHRONOLOGY AND MANAGEMENT 

This chapter contains a chronological history of the project-, an 
outline of the organizations and individuals involved in the project, the 
project budget, and the project schedule. 

2.1 Project History 

The ATLIT airplane represents the culmination of a long-term general 
aviation research program embarked on by KU-FRL in 1967. The broad goals 
of this program were to improve safety, performance, and handling qualities, 
as well as to advance the technology of the general aviation industry 
products. It has been argued that the basic control systems and aero- 
dynamic designs of general aviation airplanes have changed very little in 
as long as thirty years. Furthermore, the advanced technology which resulted 
in marked performance improvements in commercial (jet transport) and 
military aircraft had not been applied to any significant extent in general 
aviation. In addressing the goals of the research program, efforts were 
to be made to apply both existing and advanced technology to light airplane 
designs. 

Under NASA grants to FRL, the general aviation work has evolved in 
two major phases, beginning with the modified Cessna Cardinal "Redhawk" 
project (Phase I) and continuing to the present ATLIT project (Phase II). 

Phase I, the development and testing of the Redhawk, began with the" 
awarding of NASA Grant NGR 17-002-072 in 1969. The pi anform modifications 
to the Cessna C-177 Cardinal are illustrated in Figure 2.1. The changes 




Figure 2.1.- Three-view of Redhawk compared to original Cessna Model 177 Cardinal 



made to the airplane are quantified in table 2.1. The major goal in applying 
jet- transport wing technology to the Redhawk was to design a wing optimized 
for cruise efficiency with taper, increased aspect ratio, reduced area, 
and reduced thickness. The reduced wing area led to the use of high-lift 
devices to maintain takeoff and landing performance comparable to the 
unmodified airplane. The development of the Fowler and Kruger flaps for 
the Redhawk made use of two-dimensional KU wind-tunnel- test data 
(reference 7). The use of spoilers rather than ailerons for roll control 
was investigated to permit the use of full-span flaps and to provide 
flightpath control by direct-lift control- The Redhawk spoiler design also 
made use of KU wind-tunnel test data (references 8 and 9). 

First flight of the Redhawk took place in 1972. The results of the 
Redhawk performance evaluation (reference 3) show increased cruise speed 
(decreased Cpj*3), increased maximum lift coefficient, and smoother ride 



in turbulence as a by-product of reduced wing area (increased wing 

loading). The Redhawk spoilers provide adequate roll control with neither 

deadband nor nonlinearity in roll response. The lack of any significant 

net yawing moments during rolls with these spoilers makes it possible to 

make coordinated turns with no rudder deflections. The Redhawk lateral 

control forces, due to friction, are high in the all-mechanical system. 

This results, in part, from the use of cams, allowing individual movement 

of the spoilers for roll control as well as allowing symmetric spoiler 

displacements for direct-lift control. However, there is positive wheel 

centering in all flight conditions. 

Analysis of the Redhawk climb performance (reference 3) shows reduced 

maximum rate of climb in comparison with the unmodified Cardinal, as 
predicted by preliminary design analysis (reference 10). This 



TABLE 2.1 
COMPARISON OF REDHAWK AND CESSNA CARDINAL DESIGN SPECIFICATIONS 

(FROM REFERENCE 3) 



Gross Weight, K, (lb) 

Wing Area, m^, (sq ft) 

Wing Loading, N/mZ, (Ib/sq ft) 

Span, m, (ft) 

Aspect Ratio 

Taper Ratio 

Twist, deg 

Dihedral, deg 

Airfoil Section 

Inboard 

Outboard 
Trailing-edge Flap 

Type 

Span, percent 

Area (both), m^, (sq ft) 
Leading-edge Flap 

Span, percent 

Deflection 
Aileron 

Type 

Chord, percent 

Span, percent 
Spoiler 

Type 

Span 

Inboard 
Outboard 
Maximum Deflection 



Cardinal 


Redhawk 


11,120, (2500) 


11,120, (2500) 


16.23, (175) 


10.21, (110) 


648, (14.3) 


1089, (22.7) 


10.82, (35.5) 


9.58, (31.4) 


7.4 


9.0 


0.7 


0.5 


3.0 


3.0 


1.5 


3.0 


NACA 64A215 


NACA 2412 


NACA 64A212 


NACA 2409 


Single Slot 


Fowler 


53 


47 


2.74, (29.5) 


2.93, (31.5) 


- 


Kruger 


- 


83 


- 


135" 


Frlse 


Round Nose 


41 


24 


33 


36 



Modified Mitsubishi 

28.5 
32 
53* 



10 



results directly from the effect of the reduced Redhawk span (increasing 
span loading) in increasing induced drag. Calculations of reference 3 
show climb performance improvements which would have accrued due to 
construction of the Redhawk wing with the original Cardinal wing span. 

The Redhawk is presently active at the University of Kansas in a 
flight-test program to evaluate the direct-lift control spoilers as 
fllghtpath control devices on ILS approaches. 

Phase II of the general aviation research program, the development 
and testing of ATLIT, began in 1972. Much of the work done on the Redhawk 
had direct or indirect applications on ATLIT. The same type of parametric 
analysis that was performed during the design of the Redhawk indicated 
that an ATLIT wing with Increased aspect ratio, reduced area, and using the 
GA(W)-1 airfoil would improve single- and multi-engine climb performance 
and cruise performance. As on the smaller Redhawk wing, the ATLIT wing 
required some form of a high-lift device. With the confidence in roll- 
control spoilers gained on Redhawk, the application of these devices on 
ATLIT freed the full span of the wing trailing edge for use of the Fowler 
flap. 

The maiden flight of ATLIT took place on October 12, 1974, at the 
Piper Aircraft Corporation Facility in Lakeland, Florida. Mr. W. P. Kelly 
of Piper was the test pilot. Following a period of debugging, final 
construction, and about 10 hours of acceptance testing, the airplane was 
delivered by Piper to NASA-Langley Research Center on November 26, 1974. 
At LRC, the airplane was grounded until April 1975, for inspection by 
LRC Quality Assurance Office and for installation of the flight-test 
instrumentation system and recording package. Instrumented test flights 
began May 28, 1975. 



11 



Figure 2.2 presents some of the milestones during the Redhawk and 
ATLIT research projects. 

2.2 Project Support Organization 

The groups and organizations involved in the various aspects of the 
ATLIT research program are indicated in Figure 2.3. A description of the 
extent of each organization's contribution and related literature published 
is presented here. 

The Safety and Operating Problems Branch (SOPB) in the Flight Research 
Division (FRD) at NASA-LRC has had responsibility in funding the general 
aviation work done under NASA Grant N6R 17-00-072. Mr. Harold L. Crane 
(LRCJ has been the project technical monitor of this grant and dther grants 
related to the ATLIT project at Wichita State University, North Carolina 
State University, and Princeton University. He was also the LRC project 
engineer for the ATLIT flight-test program. 

The University of Kansas has been responsibile for overall ATLIT program 
management. 

Much of the associated project work was performed under subcontract 
from KU. Dr. David L. Kohlman (KU ) is the principal investigator for the 
project. Mr. Bruce J. Holmes (KU-Doctor of Engineering Degree Candidate) was 
the KU project engineer for the ATLIT flight-test program at LRC. The design 
of an advanced technology light twin-engine type of airplane was first suggested 
in reference 4. Development of a cruise-optimized planform was performed with 
the aid of KU computer programs. References 1, 11, 12, 13 and unpublished data 
resulted largely from work done by KU personnel on the ATLIT project. 

Under subcontract from KU, Wichita State University (WSU) did the 
wind tunnel development on the Fowler flap and roll-contro'l spoiler systems 

2. Ibid. 



12 



1969 1970 1971 1972 1973 1974 1975 



REDHAWK 

Award Grant 

First Flight 

Flight Demonstration 
at LRC 

Begin Flight Testing 

ATLIT 

Conceived 

GA(W)-1 Airfoil Design 

Preliminary Design 

WSU Flap and Spoiler 
Development 

First Flight 

Begin Flight Testing 



▲ 
▲ 



Figure 2.2 Milestones in General Aviation NASA Grant Research at the 
University of Kansas. 



13 



UNIVERSITY OF KANSAS 
PROJECT AAA NAGEMENT 



NASA-LANGLEY RESEARCH CENTER 
PROJECT SPONSORSHIP/SUPPORT 



WICHITA STATE UNIVERSITY 

2-D AND 3-D WIND TUNNEL 

FLAP AND SPOILER 

DEVELOPMENT 



ROBERTSON AIRCRAFT CORPj 

ATLIT DETAIL DESIGN 
SUPERCRITICAL 
PROP DESIGN 



PACIFIC PROPELLER CORP. 

SUPERCRITICAL PROP 
CONSTRUCTION 



PIPER A IRC RAFT CORP. 

ATLIT CONSTRUCTION 
MAINTENANCE SUPPORT 



NASA -WALLOPS FLIGHT CENTER 
FLIGHT-UST SUPPORT 



NORTH CAROLINA STATE 
UNIVERSITY 

COMPUTER PREDICTIONS 
OF ATLIT PERFORMANCE AND 
STABILITY AND CONTROL: 
DRAG/ POWER EXTRACTION 
TECHNIQUE 



PRINCETON UNIVERSITY 

RED NAVION SPOILER 
NONLINEAR ROLL CONTROL 
EVALUATION 



LOW TURBULENCE 
PRESSURE TUNNEL 

GA(W) AIRFOIL 
VERIFICATION 



V/STOL TUNNEL 

3-D SPOILER AND 
FLAPS ON GA(W)-1 
WING 



FULL-SCALE TUNNEL 
ATLIT TESTS (1976) 



ANALYSIS AND 

COMPUTATION 

DIVISION 

DATA REDUCTION 



Figure 2.3.- ATLIT project support organization. 



for the ATLIT wing. Dr. William H. Wentz, Jr. (WSU) had responsibility 
for this development work. In addition to the 2-D wind tunnel work on 
the airfoil-flap-spoiler configuration, VJSU ran reflection plane wind- 
tunnel tests directly under an NASA-LRC grant. The purpose of these tests 
was to document the ATLIT airfoil-flap-spoiler configuration in th^ee 
dimensions. This testing Included documentation of wing forces, spoiler 
hinge moments, and tufted stall patterns. References 14, 15, 16 and 17 are 
products of this work. 

Robertson Aircraft Corporation , Renton, Washington, under Piper Aircraft 
Corporation and KU subcontracts did a majority of the ATLIT detail design. 
They also designed a set of propellers incorporating a supercritical airfoil 
for testing on ATLIT. The ATLIT design^ drawings and design loads analysis 
were prepared by Robertson. The early preliminary design work on ATLIT was 
done by Mr. James D. Raisbeck of Robertson. After Mr. Raisbeck's 

departure from the company, Mr. John T. Calhoun had primary responsibility 

3 4 
for completing the ATLIT detail design. Unpublished reports * contain data 

from the Robertson Corporation work on ATLIT. In addition, the engineering 

design drawings for ATLIT listed in Table 3.2, Chapter 3, were prepared. 

Construction of the Robertson-designed supercritical propellers was 
done by Pacific Propeller Corporation , Kent, Washington, under subcontract 
from Robertson. 

Construction of the ATLIT wing was done under a KU subcontract to 
Piper Aircraft Corporation , Lakeland, Florida. Mr. H. Raymond Bazo (Piper) 
was the project engineer in charge of this construction. Approximately 



3. Ibid. 

4. Budish, Nathan N. : ATLIT Design Loads, Robertson Aircraft Corp. 
report TR-ATLIT-1. Prepared for the University of Kansas Center 
for Research, Inc., under MASA Grant NGR-17-002-072, June 1973. 



15 



20 hours of acceptance testing was done by Pip^r prior to delivery of the 
airplane to LRC. Piper also did approximately 5 hours of flight evaluation 
of the supercritical propellers installed on a standard PA-34 Seneca. 
Piper provides maintenance support when required for ATLIT during flight 
testing. The title to the airplane remains in Piper's name with a lease 

arrangement to KU-FRL for the purpose of flight testing at LRC. Piper work 

5 6 
related to ATLIT was documented in unpublished reports. ' 

The NASA - Wallops Flight Center , Wallops Island, Virginia., provides 
an isolated environment for flight testing. Wallops has extensive capabilities 
in flight tracking, data reduction, and ground support. These facilities 
are used for such ATLIT tests as airspeed calibrations, takeoff and landing 
performance, single-engine climb performance, and noise measurement. 

Under a grant from LRC, North Carolina State University did analytical 
work in the areas of predicting ATLIT performance and stability and control 
characteristics. Dr. Frederick 0. Smetana (NCSU) is the principle invest- 
igator for the grant. The purpose of the computer predictions of airplane 
characteristics, in addition to evaluating the ATLIT design, was to provide 
data for correlation with flight-measured characteristics and thus attempt 
to build confidence in the computer-predictive techniques. In additon to 
this work, a computer method is under development for extracting drag and 
power data from continuous, dynamic flight-maneuver data. The technique 
is presented in detail as Appendix A. Reference 18 is a product of NCSU 
work related to ATLIT. 



5. Kimberlin, Ralph D.: Flight Test Evaluation of the NASA/University of 
Kansas Advanced Technology Light Twin, Parts I and II. Piper Aircraft 
Corporation In-House Reports, 1975. 

6. Kimberlin, Ralph D.: Comparative Evaluation of the NASA/University of 
Kansas Supercritical Propellers with Standard Propellers on the PA-34-200 
Seneca I. Piper Aircraft Corporation In-House Report, 1974. 



16 



Under a grant from LRC, the Princeton University Flight Research Laboratory 

conducted an in-flight simulation to explore the effects on handling qualities 
of wind-tunnel predicted spoiler-type roll-control nonlinearities. This 

work consisted of programing a variable stability airplane for several 

different cases of nonlinearity and deadband combinations. Flight evaluations 

by LRC research pilots developed confidence that certain degrees of nonlinearity 

would be tolerable. The flight experience prepared the pilots for the possible 

cases of nonlinearity for the ATLIT first flight. Mr. David R. Ellis was 

the principal investigator, and reference 19 is a product of this grant work. 

A few months after ATLIT was conceived, the characteristics of one of 
the first computer designed airfoils, the GA(W)-1, were being documented in 
the Low-Turbulence Pressure Tunnel at LRC. Mr. Robert J. McGhee (LRC), 
working with Dr. Richard T. Whitcomb (LRC), completed development of the 
airfoil by early 1973. The airfoil, a spinoff of Dr. Whitcomb's supercritical 
airfoil work, showed promise for general aviation applications and was 
incorporated into the ATLIT design. Reference 5 is a product of this 
wind-tunnel work. 

Mr. John W. Paulson, Jr. (LRC) conducted 3-D wind-tunnel investigations 
in the LRC V/STOL Wind Tunnel on a wing with a GA(W)-1 section. The tests 
included evaluation of Fowler flaps with roll-control spoilers, and plain 
and slotted flaps with roll -control ailerons. These tests generated wing-force 
data with the three types of flap systems and data on roll -control character- 
istics with either ailerons or spoilers. References 20 and 21 are products 
of these wind-tunnel tests. 

In the fall of 1976, ATLIT will be tested in the LRC Full-Scale 
(30- by 60-Foot) Wind Tunnel . 



17 



The Analysis and Computation Division (ACD) at LRC has supported the 
project in data-reduction tasks- A sample of the work this division 
performs is illustrated in figure 2.4. The process illustrated in the 
figure traces the reduction of flight-test data from analog flight data 
on magnetic tape to the final engineering units time histories. The ATLIT 
project will continue to receive support from other organizations at LRC. 
Planned testing will involve personnel outside the Flight Research Division 
for propeller noise tests and stability derivative extraction tests. 

2.3 Project Budget 

The total funding for the ATLIT project is outlined below. Funding 
was obtained from the National Aeronautics and Space Administration, 
Langley Research Center, Hampton, Virginia, and the University of Kansas 
on a cost sharing basis. All income and outflow of project funds were 
handled through the business office of the Center for Research, Inc. 
(CRINC), by the principal investigator for the project. Table 2.2 outlines 
the project budget in terms of grant (cost-shared) funding and costs 
incurred by LRC in directly supporting the ATLIT flight- test program. The 
amounts of cost shared funding provided by KU are excluded from the table. 
These amounts generally consisted of small matching funds from the 
University for the principal investigator's salary during the academic year. 
The Langley direct funding does not include overhead charges for the 
operation of the airplane at Langley. 

Each item in the breakdown of funding in table 2.2 is underlined 
and explained below. 

The funding for ATLIT development includes conceptual design of the 
wing, stability and control analysis, handling qualities analysis, airfoil 



18 




Convert digitized 
voltages to 
engineering units 



Tabulate flight data 
in engineering units 



Plot flight data 



Figure 2.4.- Data reduction work-flow diagram. 



19 



TABLE 2.2 

TOTAL ATLIT PROJECT BUDGET 

Grant Funding (during four year period 3/72 to 6/76) 

ATLIT Development $ 245,519.00 

Wing Construction 359,000.00 

Supercritical Propeller Design and Construction 22,000.00 

Flight Test Program Support (two years) 62,498.00 

Hull and Liability Insurance 9,500.00 

(1) Total Grant Funding $ 698,517.00 

Langley Direct Funding (during two year period, 6/74 to 6/76) 
Salaries and Wages 

Engineering, Pilots, and Maintenance $ 123,000.00 

Instrumentation Support 130,000.00 

Wallops Flight Center Support 1,500.00 

Langley Chase Aircraft Support 2,500.00 

Standard Seneca Rental 350.00 

Airplane Direct Operating Costs (for 85 flight hours) 1,100.00 

Computer Time 6,000.00 

Travel 4,500.00 

Miscellaneous Equipment, Parts, and Supplies 4,500.00 

Grants (other than K U) 80,000.00 

(2) Total Langley Direct Funding $ 353,450.00 

Total Program Costs (1) + (2) $1,051,967.00 



20 



development studies (at WSU) and selection, two- and three-dimensional 
wind-tunnel development work of the Fowler flap and spoiler roll -control 
systems, and detailed engineering design. 

The funds for wing construction were paid under a subcontract from 
KU to Piper Aircraft Corporation. 

Under a subcontract from KU, funds for the supercritical propeller 
design and construction were paid to Pacific Propeller, Inc., under contract 
from Robertson. 

The flight-test program support funding was awarded primarily to pay 
for the services of one KU graduate student at LRC to serve as KU project 
engineer during the flight testing of ATLIT. 

Under the $1 per year airplane lease arrangement between CRINC and 
Piper Aircraft Corporation, hull and liability insurance was required for 
ATLIT. Funds were provided through the grant for coverage against any 
possible claim not covered under the Federal Tort Claims Act. Any claims 
involving negligence on the part of the Federal Government (NASA) would 
be covered by this act. 

Under funding for salaries and wages , the amount for engineering , 
pilots, and maintenance covers one full-time engineer (2 years), two 
one-quarter-time research pilots (1 year), one full-time mechanic (2 years), 
and a one- third-time maintenance supervisor (2 years). The instrument a tion 
support is contracted and includes salaries, wages, and company overhead for 
one engineer and one technician. 

Funds for the Wal lops F light Center support paid for about 95 man-hours 
of services during two tower- flyby airspeed-calibration flights and two 



21 



single-engine climb test flights. The services included radar tracking 
with recorded time histories and meteorological data recording. 

The Langley chase aircraft support consisted of approximately 25 hours 
of flight time in various aircraft. These flights were made to observe 
and/or photograph ATLIT during trailing anemometer airspeed calibration 
tests and tuft studies. Chase aircraft used included fixed-wing single- and 
multi-engine airplanes and helicopters. The cost for operating these chase 
aircraft was estimated to average $100 per flight hour, including ground 
and flight crew costs. 

Approximately 5 hours of flying was done in an unmodified PA- 34 with 
the funds indicated under standard Seneca rental . These flights were made 
to document performance characteristics of the standard Seneca and to do 
tuft studies. 

Airplane direct operating costs are based on fuel, oil, filters, tires, 
and miscellaneous expendable parts used during approximately 85 hours of 
research flying (from April 1975, to May 1976), This direct operating cost 
averaged about $12/f light hour. No account has been made in this analysis 
for avionics repair costs. 

The funds for digital computer time represent computer costs for work 
by both ATLIT project personnel and Analysis and Computation Division 
support personnel . 

Travel funds include all ATLIT- related travel by LRC employees with the 
exception of trips to technical society meetings. 

The amount for miscellaneous equipment, parts, and supplies , includes 
purchases of a digital fuel monitor for accurate weight control during 



22 



flight testing and a programmable pocket calculator for flight test data 
reduction. Also included is the cost of magnetic tape for the flight 
data recorder. 

The funds for grants (other than KU) include work by North Carolina 
State University on predictions of ATLIT lift, drag, moments, performance, 
and stability and control characteristics (Chapter 5) as well as work on 
a method for extracting drag and power data from dynamic maneuvering 
flight data (Appendix A). Also included is an in-flight simulator 
experiment to evaluate the influence of spoiler-type roll -control 
nonlinearities on lateral handling qualities. This work was performed 
by Princeton University. 

Table 2.3 presents the project costs which may be charged to the 
operation of the airplane during the flight test program involving about 
85 research flight hours. The result of this analysis suggests that the 
cost to Langley Research Center in operating a flight test program with 
the scope and duration of the ATLIT project is about $4,000 per flight 
hour or about $170,000 per flight-program^ year. No account is made in this 
analysis for LRC overhead costs. The salaries and wages figured into this 
average cost account for approximately six months of start-up time for 
the flight-test phase of the project, one year of active flying, and about 
six months of data analysis and report preparation. The flight hours and 
man-years used are representative of those required to document airplane 
characteristics including airspeed and angle-of-attack calibrations, 
extensive tuft studies, lateral handling qualities, stall characteristics, 
and cruise and climb performance. 



23 



TABLE 2.3 



TOTAL FLIGHT PROGRAM OPERATING COSTS FOR 85 HOURS OF RESEARCH FLYING 



Grant Funding 

Flight Test Program Support (two years) 

Hull and Liability Insurance 
Lanqley Direct Funding 

Salaries and Wages 

Wallops Flight Center Support 

Langley Chase Aircraft Support 

Standard Seneca Rental 

Airplane Direct Operating Costs (85 flight hours) 

Computer Time 

Miscellaneous Equipment, Parts, and Supplies 
Total Flight Program Operating Costs 



$ 62,498.00 
9,500.00 

$253,000.00 
1,500.00 
2,500.00 
350.00 
1,100.00 
6,000.00 
3,200.00 

$339,648.00 



Approximate cost per flight hour (for 85 research flight 

hours) $4,000/HR, 



Approximate cost per fl ight- prog ram-year 



$170,000/YR, 



24 



2.4- Project Schedule 

Since ATLIT first flew on October 12, 1975, about 130 hours of flight 
time in approximately 60 flights have been logged. Of the total flight 
time, about 85 hours have involved research work, with the remaining hours 
consisting of ferry time. Figure 2.5 presents the overall ATLIT project 
timetable. 

Following the first flight, the LRC Quality Assurance (QA) Office sent 
a representative to the Piper plant in Lakeland, Florida, for an inspection 
of the airplane prior to its delivery to NASA. These QA inspections were 
addressed solely to matters of mechanical safety of flight. Matters 
concerning operational safety of flight (handling qualities and the like) 
were taken up in NASA-LRC safety committee meetings. The outcomes of the 
QA inspections and safety committee meetings included several recommendations 
which were to be implemented prior to beginning ATLIT flight operations 
from Langley Field. 

Most of the recommendations of the two investigating groups were 
implemented before final adjustments to the airplane at the Piper plant. 
The decision was made at Langley to have the airplane delivered (on 
November 26, 1975) with a small amount of work remaining to be finished on 
the airplane. This would allow completion of final preflight test-airplane 
modifications at Langley with the QA inspection personnel readily available. 
In addition, installation of the instrumentation system and data recording 
package could begin immediately upon arrival of the airplane. 

The original planning for the flight- test program called for research 
flights to begin in early 1975, and continue to late 1975, with documentation 
of the flight-test results planned for the first 6 months of 1976. 



25 






LANGLEY RESEARCH CENTER 








MILESTONES 


iq74 


l<i7R 


iq7fi 












J J.A.S.O.N.D 


J,F,M,A,M,J,J,A.S,0,N,D 


J,F,M,A,M,J,J 




I 


KU Pro,iect Manager to LRC 


£ 


1 


1 


1 


1 




1 




_ 




— 


— 






_ 


_ 


_ 




— 




_ 




„ 


— 




„ 








— 






— 






— 






r 


— 


— - 






3 


ATLIT First Flight 


r 


■ 


1 


■ 


I 


■ 


1 


■ 


■ 

▲ 


U 


■ 


I 


■ 

■ 




- 




— 








- 


- 






- 




- 




— 


— 




- 


— - 





— 






- 






— 








— 


— - 






4 


Piper Acceptance Tests 
































































































5 


ATLIT Delivery to LRC 




















▲ 










































































6 


Final LRC-QA Inspection 
























I^^^^H 












































[ 














1 




7 


Instrumentation Work 






1 
















^^^^^ 




























1 




1 ; 1 




1 






S 


Instrument Check Flights 






1 ■ 1 






































1 r 




■ I 1 


1 1 




9 


Airspeed Calibration 


1 ! : 














■■i^iHi 


1 














1 


i 








: ! i ! ' ' 1 




10 


Stall Characteristics 




i , 
































, [ 1 


\ ', \ [ ': 1 i 1 t 




11 


Tuft Studies 






' 1 ' ' ' 




















1 












12 


Fliqht Data for NCSU 
























■ 1 ' 


i 






13 


Spoiler Roll Performance 


■■ ! ' 










1 ' 


Milt ■ 




^^^^^^1 1 ' I ' ^ 




14 


Speed-Power Performance 


-r- ■ 1 






1 


■ 


1 : 1 1 ■ ^ , 1 1 






HppmP^^^^^^^^^^H^^! 




15 


a - Calibration 




"T" 




: 1 

1 


. , t - ; , 








16 


Sawtooth Climb Performanci 




! ' ; 


1 


: : 










1 : ' M i 

Mill 


WWHWC^^ 




17 


Spoiler Leakage Sealed 


: . ; ; . ' : ' i , ' i . ■ ■ ■ ■ ' 


' ft ' ' 1 ' ' 


' ' 1 ' ■ 1"' j' 








18 


Seneca-ATLIT Fly-Off 


■ ■ i ■ ,' j , 1 . ■ ' 1 1 ' t 


; ^ii ' 1 


' 1 ' ■ : ' ! 




r 1 




19 


Spoiler Float Fixed 


■ ; . ■ ' : ; ; ■ : ^ ■ . . i i a ^ 


^^^^^M 








20 


Interference Drag Cleanup 


I 


*-. 




^^ICT 




1 1 




21 


Stability Testinq 


1 i : 


; ] , 1 1 1 ' ! ; ■: 1 j 1 1 1 1 1 ' ' ' 1 1 1 [ 1 ' ; ! ; 








22 


Baseline Propeller Tests 


' j ; , ■ ' ■ ■ , ; . 1 ; : i ■ : ; 1 ' 1 j 1 ' ' '';'■!!' \ \ 
■ 1 ■ • 1 ' 'II''' ■ ■ 1 1 1 ' 1 ■ 1 1 ' 1 1 1 




1 i 1 




it 




23 
24 


Supercritical Prop. Tests 




1 




■ i i 


1 


1 




Full-Scale Wind Tunnel^ 










\ 




25 


Final Fliqht Evaluation^ 


1 1 1 , 1 1 1 i 1 : i 1 ! 1 , 1 ; 1 , 


L_ 


i 1 1 1 , 1 . 1 , 1 : ; ' ! : 1 1 ! ; ! ! 


i 1 




1 




NQ 


' Testing to begin in November, 1976. 
^ In wind-tunnel optimized configuration 







Figure 2.5.- ATLIT flight- test program timetable. 



After arrival of the airplane at LRC, the work required to meet QA office 
standards combined with instrumentation difficulties to delay the first 
instrumented test flight until May 19, 1975. Further setbacks to the 
planned flight program were encountered upon the discovery of (and research 
to cure) a region of wing/body interference- induced flow separation at 
climb speeds. An estimated four months was spent investigating this flow 
problem. An estimated three months was invested in attempts to refine 
the accuracies of the flight data instrumentation system. One month of 
time was lost due to defective recording tape for the flight data 
recording system. These developments necessitated a twelve month 
extension (May 15, 1976 to May 15, 1977) of the grant for the purpose 
of fully documenting the airplane cruise and climb performance, and to 
document the characteristics of the supercritical propellers. The lease 
arrangement for the airplane has also been extended. The extension will- 
also allow for the planned Full-Scale (30- by 60-Foot) Wind-Tunnel tests. 



27 



CHAPTER 3 

AIRPLANE MODIFICATIONS 

This chapter presents a detailed description and the design approach 
for the ATLIT wing and supercritical propellers. In presenting the 
airplane details, comparisons are made with the unmodified Piper PA- 34-200 
Seneca I wing and standard propel lers- 

3.1 General Seneca/ATLIT Description and Comparison 

The PA-34 Seneca I is representative of general aviation light 
twin-engine airplanes which are used extensively by third-level air carrier, 
air taxi, corporate, and private operators. It is a low-wing airplane 
with retractable landing gear and a maximum range of 745 n.m. at a 
75-percent power cruise speed of 162 knots. With a gross weight of 
1.87 kN (4200 lb), the airplane seats up to seven occupants. The power 
plants are normally aspirated, reciprocating engines with constant-speed 
propellers. 

The selection of the Seneca I for the project modifications followed 
a major goal in the ATLIT design, that of improving single-engine climb 
performance. General aviation airplanes in the light (less than 26.69 kN 
(6000 lb)), propeller-driven, twin-engine (normally aspirated), four- to 
eight-passenger class are virtually all very limited in single-engine 
climb performance at gross weight. For ten airplanes of this class on 
the market in 1975-1976, single-engine rates-of-climb at sea-level average 
96 m/min (320 fpm) and single-engine service ceilings average 2000 m (6600 ft) 



28 



The Seneca I, with a single-engine rate-of-climb (at sea level and 
gross weight) of 57 m/min (190 fpm) and a single-engine service ceiling 
of 110 m (3650 ft), is a typical example. Reference 2Z includes comments 
that, short of significant reduction in payload or range, no technology 
has been developed to improve climb performance. Increases >n horsepower 

are Gconomically unacceptable. Even turbo/supercharging, while, on the 
average, doubling the single-engine service ceilings, does not improve the 
sea-level rates-of-cl imb. 

The conceptual studies which led to the ATLIT design revealed the 
potential of these approaches to improving engine-out climb: 

1. Planform changes for lower induced drag with high-lift^large 
span flaps and roll -control spoilers. 

2. GA(W)-1 airfoil for higher L/D, especially at climb, and for 

higher C. 

max 

3. Supercritical propellers designed for increased propulsive 
efficiency. 

Figure 3.1 compares the planform, flap, and lateral control 
modifications of ATLIT with the unmodified Seneca. Pertinent dimensions 
for both airplanes are presented in table 3.1. 
3.2 ATLIT Design Description 

To supplement the following detailed description of ATLIT, a three-view 
is presented as figure 3.2 and a list of all engineering design drawings 
for ATLIT is given in table 3.2. 



29 



FOWLER FLAP EXTENDED 
SPOILER 



CO 

o 




Figure 3.1.- Planview comparison of the ATL.IT and the unmodified Piper PA-34-200 Seneca I 



TABLE 3.1 - COMPARISON OF ATLIT AND PIPER PA-34-200 SENECA I DESIGN SPECIFICATIONS 

ITEM ATLIT PA -34 

GROSS WEIGHT, N (lb) 18700 (4200) 18700 (4200) 

WING AREA, M^ (sq ft) 14.4 (155.0) 19.4 (208.7) 

TAPER RATIO 0.5 1.0 

ASPECT RATIO 10.32 7.25 

SPAN, (V\ (ft) 12.19(40.0) 11.85(38.88) 

SPAN LOADING. N/M (lb/ft) 1 536 (105) 1 581 (108.2) 

WING LOADING, N/M (Ib/sq ft) 1298 (27.1) 964 (20.12) 

FLAP TYPE FOWLER PLAIN 

SPAN, PERCENT OF b 88 50 



CHORD. PERCENT OF c 30 20 

TRIANGULAR 
S PERCENT OF b/2 ^^OS^S^' SECTION 

CHORD, M (inches) q Qgg .^ ., 



AIRFOIL GA(W)-1 {17%t/c) eS^-AlS 



CO 





1.93 m 



NORMAL GROUND LINE 



Figure 3.2.- ATLIT three-view. 



TABLE 3.2- ATLIT ENGINEERING DESIGN DRAWINGS 



CO 
GO 



DRAHISG NO. 

90-000140 
90-000145 
90-010010 
90-010011 
90-020000 
90-100000 
90-110000 
90-110001 
90-110002 
90-110100 
90-110101 
90-110102 
90-110106 
90-110107 
90-110108 
90-110200 
90-110210 
90-110211 
90-110212 
90-110250 
90-110255 
90-110300 
90-110301 
90-110302 
90-110400 
90-110401 
90-110402 
90-110403 
90-110404 
90-110500 
90-110550 
90-110555 
90-110600 
90-110700 
90-110701 
90-110702 
90-110703 
90-110704 
90-110710 
90-110711 
90-110712 
90-110713 
90-110714 
90-110715 
90-110716 
90-110717 
90-110718 



DRAWING W. 



Lines - Nacelle 


90-110719 


Lines - Wing Tip 


90-110270 


Master Diagram - Hing 


90-110721 


ATLIT Wing General Dimensional Info. 


90-110722 


ATLIT Three .View Drawing 


90-110723 


Wing Installation 


90-110724 


Hirg Assy Complete 


90-110728 


Wing Assy Outboard 


90-110729 


Wing Assy Inboard 


90-110730 


Spar Assy - King Main 


90-U0731 


Cap - Wing Main Spar Lower 


90-110732 


Cap - Wing Main Spar Upper 


90-110733 


Doubler Instl . Brake Clearance 


90-110734 


Side Brace Supt Instl. - Landing Gear 


90-110735 


Fitting - Side Brace - Landing Gear 


90-110736 


Spar Assy - Wing Rear (64%) Outbd 


90-110737 


Cap 64J Spar - Outbd 


90-110738 


Tee 54i Spar - Outbd 


90-110743 


Angle 641 Spar Splice 


90-110744 


Spar Assy - Wing (64%) Inbd 


90-110745 


Cap- 64* Spar - Inbd 


90-110748 


Sub Spar Assy - 24% - Wing 


90-110749 


Cap - 24% Spar - Details 


90-110750 


Angle - 24% Spar Splice 


90-110751 


Spar Assy - Center Section, Wing 


90-11075? 


Cap - Lower, Center Section, Wing 


90-110753 


Cap - Upper, Center Section, Wing 


90-110754 


Tie - Lower, Center Section, Wing 


90-110756 


Tie - Upper, Center Section, Wing 


90-111001 


Stringer - Wing 


90-111002 


Stringer Assy - 24% 


90-111003 


Angle - Stringer Splice 


90-111004 


Door - Wing Access 


90-111005 


T. E. Instl. 


90-111006 


Flap Track Instl. - Sta. 28.00 


90-111007 


Flap Track Instl. - Sta. 100.00 


90-111008 


Flap Track Instl. - Sta. 171.00 


90-111009 


Flap Track Instl. - Sta. 231.00 


90-111010 


Track - Wing Flap 


90-iiion 


Sib Instl - Sta. 45.50 Wing T. E. 


90-111012 


Rib Instl - Sta. 64 Wing T, E. 


90-111013 


Rib Instl - Sta. 79.50 Wing T. E. 


90-111014 


Rib Inst! - Sta. 114.00 Winq T. E. 


90-111015 


Rib Instl - Sta. 126.00 Wing T. E. 


90-111020 


Rib Instl - Sta. 142.50 Wing T. E. 


90-111021 


Rib Instl - Sta. 151.50 Wing T. E. 


90-111022 


Rib Instl - Sta. 186.00 Wing T. E. 


90-111023 




90-111024 



TITLE 

Rib Instl - Sta. 201.00 Wing T. E. 

Rib Instl - Sta. 215.00 Wing T. E. 

Beam Wing T. E. 

Angles - Wing T. E. 

Angle - T. E. 

Plate - Wing T. E. 

Zee Section - T. E. Rib 142.50 

Doubler - T. E. Rib 142.50 

Clip - T. E. Rib 

Angle - K. S. 114.00 

Zee Section - T. E. Rib W. S. 114.00 

Channel - H- S. 114.50 

Angle - W. S. 126.00 

Zee Section - T. E. Rib 126.00 

Channel - T. E. Rib 126.00 

Zee Section - T. E. Rib 186.00 

Channel - T. E. Rib 186.00 

Angle - T. E. Rib 151.50 

Zee Section - T. E. Rib 151.50 

Channel - T. E. Rib 151.50 

Angle - T. E. Rib 215.00 

Channel - T. E. Rib 215.00 

Zee Section - T. E. Rib 215.00 

Angle - T. E. Rib 201.00 

Zee Section - T. E. Rib 201.00 

Channel - T. E. Rib 201.00 

Angle - Wing R. E. 

Angle - Track W. S. 28.00 

Rib Instl - Wing Cant. - Sta. 28 

Rib Instl - Wing - Sta. 41 

Rib Instl - Wing - Sta. 54 

Rib Instl - Wing - Sta. 64 

Rib Inst? - Wing - Sta. 86.00 

Rib Instl - Wing - Sta. 100 

Rib Instl - Wing - Sta. 111.00 

Rib Assy - Wing - Sta. 126 

Rib Assy - Wing - Sta. 141 

Rib Assy - Wing - Sta. 156 

Rib Assy - Wing - Sta. 171 

Rib Assy --Wing - Sta. 186 

Rib Assy - Wing - Sta. 201 

Rib Assy - Wing - Sta. 216 

Rib Assy - Wing - Sta. 231 

Rib - Wing Cant - Sta. 28 

Fitting - Wing Mount, Fwd 

Fitting - Wing Mount, Aft 

Tee - Rib, Cant. - Sta. 28 

Angle - Rib, Cant. - Sta. 28 



TABLE 3.2- Continued 



CO 



ORAWIHG HO. 

90-111025 
90-111026 
90-111027 
90-111028 
90-120000 
90-120100 
90-120101 
90-120102 
90-120103 
90-120104 
90-120105 
90-120106 
90-120107 
90-120108 
90-120109 
90-120111 
90-120114 
90-120115 
90-120116 
90-120117 
90-120200 
90-120201 
90-120202 
90-120203 
90-120204 
90-120205 
90-120206 
90-120207 
90-120208 
90-120209 
90-120210 
90-120300 
90-120301 
90-120302 
90-120303 
90-120304 
90-120305 
90-120306 
90-120307 
90-120308 
90-120309 
90-120310 
90-120311 
90-120315 
90-120400 
90-120532 
90-120600 
90-120700 



TITLE 

Frame Assy - Leading Edge, Inbd Wing 

Frame Instl. - Canted, Inbd Wing 

Frame - Leading Edge, Inbd Wing 

Angle - Leading Edge, Inbd Wing 

Flap Instl. 

Flap Assy - Inbd 

Carriage Assy. W. S. 28,00 

Rib Assy - Flap W. S. 42.40 

Rib Assy - Flap W. S. 56.80 

Rib Assy - Flap W. S. 71.20 

Rib Assy - Flap W. S. 85.60 

Rib Assy - Flap W. S. 28.00 

Carriage - Flap W. S. 28.00 

Gusset - Inbd Flap 

Gusset - Inbd Flap 

Spar Assy - Flap Inbd 

Clip - Inbd Flap 

Clip - Inbd Flap 

Clip - Inbd Flap 

Clip - Inbd Flap 

Flap - Assembly Center 

Carriage - Flap H. S. 100. Oa 

Rib Assy - Flap H. S. 100.00 

Carriage Assy - W. S. 100.00 

Rib - Flap H. S. 114.20 

Rib - Flap W. S. 128.40 

Rib - Flap W. S. 142.60 

Rib - Flap W. S. 156.80 

Gusset - Flap Drive W. S. 165.00 

Spar Assy - Flap Ctr. 

Angle - Center Flap 

Flap Assembly - Outboard 

Carriage - Flap W. S. 231.00 

Rib Assy U. S. 230.34 

Carriage Assy W. S. 231.00 

Rib - Flap H. S. 186.00 

Rib - Flap H. S. 201.00 

Rib - Flap H. S. 216.00 

Carriage - Flap W. S. 171.00 

Rib Assy - Flap W. S. 171.00 

Carriage Assy W. S. 171.00 

Spar Assy Outbd Flap 

Gusset - Flap Drive - 227 

Angle - Clip Flap Spar 

Pushrod Assy - Wing Flap 

Bracket - Flap Lever 

Support Bracket - Flap B/C 

Lever Assy - Flap 



90-120800 
90-120801 
90-120802 
90-120803 
90-120804 
90-120805 
90-120806 
90-120900 
90-130000 
90-130100 
90-130200 
90-130300 
90-130401 
90-130402 
90-130403 
90-130404 
90-130405 
90-130406 
90-130407 
90-130408 
90-130409 
90-130410 
90-130411 
90-130501 
90-130503 
90-130504 
90-130505 
90-130506 
90-130507 
90-130508 
90-140000 
90-140001 
90-140002 
90-140003 
90-140004 
90-140005 
90-140006 
90-140007 
90-140008 
90-140009 
90-140011 
90-140012 
90-140013 
90-140014 
90-140015 
90-140016 
90-140017 



Inbd Flap 
Inbd Flap 
Weld Assy 
Weld Assy, 227.0 
Flap Bracket Attachment 



TITLE 

Link Assy - Flap Drive 

Universal - Flap Drive 

Bracket 

Bracket 

Bracket 

Bracket 

Bracket 

Angles - 

Spoiler Instl. 

Spoiler Assy 

Spoiler Assy Center 

Spoiler Assy - Outbd 

Pushrod - Spoiler 

Link Assy - Spoiler 

Lever Assy - Spoiler 

Sector Assy - Spoiler Drive 

Bracket Assy - Spoiler 

Hinge Assy - Spoiler 

Hinge Half Assy - Spoiler 

Hinge Half - Spoiler 

Fitting - Link Attachment 

Bracket Details - Spoiler 

Rod End - Spoiler Drive 

Bracket Assy - Spoiler Pulley 

Spring - Spoiler 

Bracket Assy - Spoiler 

Bracket - Details - Spoiler 

Bracket Assy - Spoiler 

Bracket Details - Spoiler 

Cable Assy - Spoiler 

Nacelle Instl. 

Rib Instl. - Nacelle - Wing Sta. 64.00 

Rib Instl. - Nacelle - Wing Sta. 86. CO 

Frame Instl. - Nacelle - Wing Sta. 82.26 

Frame Instl. - Nacelle - Wing Sta. 94.26 

Frame - Nacelle - Wing Sta. 82.25 

Frame Instl. - Nacelle 64* 

Bracket - Nacelle 

Angle - Nacelle - Rib 

Angle - Nacelle - Rib 

Frame - Nacelle - Sta. 94.26 

Frame - Nacelle - Sta. 94.25 

Frame - Nacelle - Sta. 94.25 

Fitting - Upper, Nacelle 

Fitting - Lower, Nacelle 

Tee - Nacelle Rib 

Angle - Nacelle Rib 



TABLE 3.2- Concluded 



DRAWING HO. 



TITLE 



DRAWING NO. 



TITLE 






90-400100 
90-400101 
90-100102 
90-400103 
90-400104 
90-400105 
90-400106 
90-400107 
90-400108 
90-400109 
90-400110 
90-400111 
90-400112 
90-400113 
90-400114 
90-400115 
90-400115 
90-400117 
90-400118 
90-400119 
90-400120 
90-400121 
90-400122 
90-400123 
90-400124 
90-400125 
90-400126 
90-400127 
90-400128 
90-400129 
90-400130 
90-400131 
90-500000 
90-600000 
90-700100 
90-800100 
90-800101 
90-800110 
90-800111 
90-800211 
90-800212 
90-800213 
90-800214 
90-800215 
90-800215 
90-800217 
90-800218 



Fuselage Structure Assy 

Frame Instl. - F. S. 73.04 

Frame Instl. - Sta. 106.628 

Fitting - Fwd Fuselage Wing Attach 

Fitting - Wing Mount, Aft, Fuselage 

Gusset - Lower Fuselage - Sta. 73.04 

Gusset - Lower Fuselage - Sta. 73.04 

Side Frame - Sta. 73.04 

Channel - F. S. 73.04 

Channel - Sta. 74.105 

Frame - F. S. 73.04 

Channel - Lower Fuselage, Sta. 73.04 

Clip - Lower Fuselage, Sta. 73.04 

Bracket - Lower Fuselage, Sta. 73.04 

Web - Upper Coc)<pit - Left Side 

Web - Upper Cockpit - Right Side 

Dblr - Cockpit - Fwd - Lwr 

Channel - Cockpit - Fwd - Lwr 

Support Fittings BHO 106.628 

Dblr - Cockpit - Lwr - Aft 

Angle - Lower Fuselage - Sta. 128.737 

Clip - Lower Fuselage - Sta. 128.737 

Support Fittings BHD 106.628 

Channel - Spar Box - Left Forward 

Channel - Lwr Fuselage - Sta. 104 

Bracket - Lwr Fuselage - Sta. 106.628 

Bracket - Lwr Fuselage - Sta. 106.628 

Channel - Lwr Fuselage - Sta. 106.628 

Strap - Lwr Fuselage - Sta. 106.628 

Web - Sta. 72.105 

Channel - Lwr Fuselage - Sta. 97 

Plate - Lwr Fuselage - Sta. 97 

Wing £lec. Harness Instl. 

Hydraulic Systems Instl. 

Sender Instl . - Fuel 

Landing Gear Instl. - Main 

Gear Assy - Main 

fitting Assy Gear Attach, Aft 

Fitting Assy Gear Attach, Fwd 

Over Center Assy 

Link Assy - Over Center 

Lever Assy - Over Center 

Side Brace Assy 

Bolt Over Center Adjust 

Bracket Assy - Over Center - Spring 

Spring - Over Center - Gear 

Bracket Assy Cylinder Support 



90-800300 
90-800310 
90-800311 
90-800312 
90-800353 
90-800354 
90-800382 
90-800400 
90-800500 
90-800501 
90-800502 
90-800600 
90-900000 
90-920000 
90-920500 
90-930000 



Trunnion Assy 

Cylinder - Main Landing Gear 

Beam - Trunnion 

Brace - Trunnion 

Orifice Weld Assy 

Orifice Tube Assy 

Plate Assy - Orifice 

Bearing - Landing Gear 

Stop Instl. - Landing Gear 

Stop Assy - Landing Gear 

Up Limit Switch Instl. - Main Landing Gear 

Door Instl. - Main Landing Gear 

Fuel System Instl. 

Controls Instl, - Engine 

Lines and Tachometer Instl - Engine Instruments 

Pitot Boom Instl. 



3.2.1 ATLIT Planform Changes 

The ATLIT planform has an aspect ratio of 10.32, taper ratio of 0.5, 
and a wing span of 12.1 m (40 ft). These planform changes should produce 
performance and ride-quality improvements for the following reasons: 

1. The induced-drag term of the airplane wing may be written as 

D. = ih ^^ . (3.1) 

It follows that (for a fixed weight and velocity) the induced drag is 

D^. - -^ (3.2) 

b -e 

or the equivalent expression 

D. ~ _J^ . (3.3) 

^ S-A-e 

On ATLIT, the increase in aspect ratio offsets the decrease in wing 
area so that the product, S'A, is increased about 5 percent. As relation 
(3.3) indicates, an increase in S-A will reduce the induced drag by about 
5 percent for ATLIT. In simpler terms, the same induced-drag change may 
be explained by noting the effect in relation (3.2) of the slight increase 
in the ATLIT span over the Seneca span. In this case, increasing aspect 
ratio (alone) does not reduce induced drag; but, increasing span does. 

The effect of the 5-percent reduction in D. for ATLIT, on total 
airplane drag is small. Induced drag on ATLIT is 10 percent of total drag 
at cruise and 50 percent of total drag at climb. These contributions of 
the ATLIT planform changes to reducing total airplane drag become less 
than 1 percent at cruise and less than 3 percent at climb. 

2. With the 25-percent smaller wing area, the profile drag of the 

airplane will be reduced. As shown below, this reduction occurs in spite 

of the higher C . of the 17-percent thick airfoil on ATLIT because 
0, min 



36 



the product) C-. . S, decreases compared to the standard wing. The 



magnitude of this change is estimated, using section data, as follows: 



Airplane 


Airfoil 

17%t/c. GA(W)-1 
662 415 


0, min 
(fixed transition. 
RN = 6 X 10^) 

0.0106 

0.0101 


S 


Wing 
profile 

c . . s ^''^a 
0, min A% 


ATLIT 
PA- 34 


14.4 m^ 
19.4 m^ 


0.15 -25% 
0.20 



Assuming that wing-profile drag is about 35 percent of total airplane drag 
at cruise and about 20 percent of total airplane drag at climb, the 
28-percent change in wing-profile drag reduces airplane drag by the 
following amounts: 

change in wing-profile drag change in 
wing-profile drag X total airplane drag = total airplane 
for smaller wing drag 

at cruise 28% X 35% = 9.8% 

at climb 28%- X 20% = 5.6% 

3. The tapered ATLIT wing contributes to increasing the wing-span 

efficiency factor (e) by 3.4 percent (reference 23). However, it is 

difficult to translate this into an effect on total airplane drag without 

knowledge of the interference effects of the modified wing on the standard 

2 
fuselage. Interference-induced separation drag, varying with C. ^ will 

appear as reduced span efficiency. It is also difficult to account for 
the effect on span efficiency of reducing the wing area with no change in 
fuselage or engine nacelles. With the fuselage and nacelle wetted areas 



37 



becoming proportionately larger in relation to wing area, the effect of 
these bodies on span loading may also become proportionately larger. 
The result could be reduced span efficiency. 

Assuming no detrimental interference effects, the contribution of 
taper to increased span factor could reduce both climb and cruise drag 
by about 4 percent.* 

4. The 25-percent reduction in wing area on ATLIT will reduce the 
airplane cruise gust response by about 20 percent.** The improvement in 
ride quality (sorely needed in general aviation airplanes) would add to the 
attraction of designing light airplanes with higher wing loadings. 

The wings for both the basic Seneca and ATLIT have 3 degrees of twist 
for desirable stall characteristics. 



* 



_ C 2 q S 

At cruise, D^. - J= 9^— = lo percent of total airplane drag. Assume e 

TTeA P 

increases from 0.75 to 0.783, then for V = 170 knots and S = 14.4 m , 
cruise D^- decreases from 250 N (56 lb) to 239 N (54 lb). For climb, 

V = 90 knots, D^. decreases from 779 N (175 lb) to 745 N (167 lb). 

- n^=CL^q, S 

at cruise 

for ATLIT: n^ ^ = (0.100 deg'^) X (4.26 kPa) x(l4.4 m^) 

for PA-34: n ^ = (0.080 deg""^) X (4.26 kPa) X (19.4 m^) 

Ob 5 O 



n c " n A 
a,S a,A 

n c 



X 100 == 20% 



38 



Summarizing, the net effect of the ATLIT wing-planform changes are 
given below: 

1. Increased span for lower induced drag. 

2. Tapered wing for increased span efficiency. 

3. Reduced wing area for lower wing profile drag. 
3.2.2 Roll Control Spoilers 

The decision to provide roll control on ATLIT by means of spoilers 
freed the entire trailing edge of the wing for a high-lift device. The 
requirement for a large-span, high-lift flap follows the decision to 
design the planform with high wing loading. 

The spoilers on ATLIT can be described as vented, gapped, upper-surface, 
roll -control spoilers. Figure 3.3 illustrates the geometry of this spoiler 
in cross section. In the literature (references 24 and 25), this type 
of spoiler has been referred to as a "slot-lip aileron." In order to make 
the difference between spoilers and ailerons distinct, any reference in this 
report to roll-control spoilers will imply an aerodynamic device which 
creates airplane rolling motion by the mechanism of separated flow on only 
one wing at a time. An aileron roll -control system, on the other hand, 
creates airplane rolling motion by changing lift on both wings simultaneously 
(by deflecting the wake due to a change in camber). 

The design details of the spoilers on ATLIT are presented here. 

Spoiler Vent-Path 
In the past, typical defects of roll -control spoilers have included 
nonlinear rolling moment variations with control -wheel deflections, 
control reversal for small deflections, and reduced effectiveness at 



39 



)«4 



Spoilers deflected to nominal maximum 

I .0166 radius 
.64 c 




Rear spar 



Spoiler hinge pin 



Spoilers rigged down 



Spoilers rigged flush 

0005 forward and aft spoiler gaps 
Wing skin 



Figure 3.3.- Spoiler installati 



on detail. 



high angle of attack. Wind-tunnel spoiler-development tests (references 15, 
16, and 17) confirmed these defects for an unvented spoiler in the flap- 
deflected configurations. In general, roll-control spoilers will not 
exhibit these characteristics in a flaps-nested configuration. 

The three undesirable traits of spoilers noted above are created by 
flow conditions at the spoiler. With a Fowler flap deflected, flow 
through the flap slot is accelerated, generating additional accelerated 
flow by a "jet effect" in the region by the spoilers. Under these 
conditions, a small spoiler deflection will create flow separation; but, 
because of the higher local dynamic pressure, this separated flow behind 
the spoiler is prone to reattach itself to the wing upper surface before 
reaching the trailing edge. The net effect of this small spoiler deflection, 
then, is an increase in effective camber. The increased camber creates 
lift and results in a control reversal. With flaps up, however, flow 
conditions at the spoiler location have relatively lower dynamic pressure. 
Under these conditions, flow separation caused by small spoiler deflections 
tend to remain separated into the wake, thus creating a proper rolling 
moment. 

The preceding wind-tunnel development work was done, in part, to 
optimize the vent-path or spoiler-cavity geometry (see figure 3,3) to 
provide some relief from the objectionable spoiler traits. The vent path 
on ATLIT eliminates control reversal and reduces the nonlinearity in roll 
response. The shape of the spoiler cavity also influences the hysteresis 
characteristics in rolling moments due to spoiler deflections (reference 16). 



41 



Spoiler Leading Edge Gap 

An important development in the application of roll -control spoilers 
to airplanes was the addition of a leading edge gap to the device (see 
figure 3.3). This feature improved the linearity of hinge moments, 
improved roll -control effectiveness of small spoiler deflections, and 
reduced the lag in roll acceleration (reference 13). The first spoiler 
design of note to fly incorporating leading edge gap was the Mitusbishi MU-2 
of the early 1960's. 

Spoiler Cross-Sectional Shape 

Wind-tunnel studies (reference 17) indicate that spoiler hinge moments 
are apparently influenced by the combination of the underside contour of 
the spoiler and the leading edge gap. The cross-sectional shape of the 
spoiler apparently has only a slight effect on rolling moment charactistics. 

The triangular cross section for the ATLIT spoilers was chosen for 
its light weight and simple construction. Hinge-moment characteristics 
for this spoiler were documented in 3-D wind-tunnel tests (reference 17). 

Design for Lateral -Control Feel* 

During the mechanical design of the ATLIT lateral -control system, 
there were virtually no data available on spoiler-hinge moments. The 
mechanical system, was designed using available split-flap data to estimate 
hinge-moment characteristics of the spoilers. Figure 3.4 presents a 
comparison of the estimated and wind-tunnel measured spoiler-hinge moments. 



*Mr. John T. Calhoun of Robertson Aircraft Corporation should be credited 
with the design of the ATLIT spoiler roll -control system. 



42 



Downgoing spoiler 
gives less wheel 
moment centering 
than predicted 



-5 



0.04 



O 0^ = 0*^ 

D 4^ 
A oO 



3-D wind-tunnel 
data (Ref. 17) 

Prel i mi nary 
design estimate 
(J. Calhoun, 
Robertson Aircraft 
Corp.) 




"~i 1 r~ 

10 15 20 



30 



40 



60 



6^, deg 



a) Flaps nested 
Figure 3.4.- Comparison of predicted and wind-tunnel measured (Ref. 17) 
spoiler hinge moments for ATLIT 



43 



T 



1.0- 



Downgoing spoiler 
gives less wheel 
centering moment 
than predicted 



Preliminary design estimate 
(J. Calhoun, Robertson Aircraft 
Corp,) 



O 
□ 

A 




3-D wind- 
tunnel data 
(Ref. 17) 



Ah 
C 



0.10 



0.12 



Upgoing spoiler . 
gives more wheel \ 
decentering moment 
than predicted 



5 


— 1 — 
10 


— 1 — 
15 


20 


^s' 


30 
deg 



40 



b) Flaps 40*^ 
Figure 3.4.- Concluded. 



60 



44 



Wheel forces In a lateral -control system are tailored to provide 
wheel centering without excessive force for maximum wheel travel. Meeting 
this requirement with spoilers presents a problem, because the hinge 
moments are in the wrong direction during the first 20° to 30°* of 
deflection. 

Vented spoilers in the neutral position are subject to positive or 
opening hinge moments. As the spoiler is deflected up to some intermediate 
position, the hinge moment will go to zero, change sign, and the spoiler 
will experience a closing moment. The ATLIT spoiler-control linkages are 
designed to provide a wheel centering force during the initial spoiler 
deflections where the aerodynamic forces tend to open the spoiler and 
decenter the wheel . 

The source of the wheel centering force on ATLIT and on several other 
mechanical spoiler designs is the aerodynamically inactive spoiler. On 
ATLIT, the left and right spoilers are connected by a high-differential 
solid linkage (no cams). Figure 3.5 illustrates the gearing of the 
spoiler motion with varying wheel position. The figure shows that when 
one spoiler goes up, the other moves down, then back up slightly as the 
wheel reaches full travel. The system is designed to allow the hinge 
moments of the downgoing spoiler to oppose the wheel decentering hinge 
moments of the upgoing one (see figure 3.6). 

A comparison of estimated and measured hinge moments in figure 3.4 
indicated that ATLIT would have undesirable lateral-feel characteristics. 
In the flap-nested case, the wind-tunnel measured opening hinge moment 



45 



05 



Left spoiler up 



Right spoiler up 




Left spoiler down 
Right spoiler down 

Figure 3.5.- Gearing between spoiler deflection and control -wheel position. 




Forces due to opening spoiler hinge moments 



6 = nO 
S 



-^■aa-:-aK5yg«a:<-w^^ 



Opposing forces balance 

each other with wheel centered 



kJ-^ 



to control wheel 




0" < + <? 





Displaced spoiler ^^^„^ 
creates uribalanced \ 
wheel centering forces 



i^^l^. 




i 1 

to control wheel 



Figure 3.6.- The effect of opening (positive) spoiler hin 



)ge moments on wheel forces 



47 



of a downgoing spoiler is about one-half of the design estimate. The 
flaps-up wheel centering force was reduced accordingly. This means that 
the opening hinge moment of the upgoing spoiler dominates, resulting in 
net wheel decentering forces. The same holds true for the flaps-down 
case, except that the difference and forces involved are larger. 

The amount of friction in the ATIT lateral-control system is large. 
This friction, by itself objectionable, partially masks the aerodynamic 
wheel decentering forces for small or large control inputs. 

A conclusion which may be drawn from this analysis is that accurate 
hinge-moment data are prerequisite to the design of mechanically-actuated 
roll-control spoilers with tolerable lateral control wheel-force 
characteristics. This requirement necessitates strong justification for 
putting a spoiler system on an airplane (e.g., a need for full-span flaps 
or direct-lift control). When a strong case can not be made for spoilers, 
the relative simplicity and lower cost of an aileron roll -control system 
will sway the decision to ailerons. 

Spoiler Leakpath Seals 

As discussed earlier, there is no requirement for lower- to 
upper-wing surface venting with flaps up. Quite to the contrary, any 
leakage of pressure through the wing is to be avoided. The ATLIT design 
did not consider the effect of allowing leakage through the spoilers. 
Wind-tunnel studies late in the design stages (reference 17) indicated 
that there would be lift and drag penalties due to leakage. Therefore, 



48 



as an operational solution to the leakage effects, a rubber weatherstripping 
seal was added (see figure 3.7). As shown in the wind tunnel, a beneficial 
effect of sealing the leakpath is that the flaps-up spoiler hinge moments 
are eliminated (spoilers neutral). 

Spoiler Rigging 

With the spoiler leakpath unsealed, the spoilers floated up about 2 or 
3 degrees. This floating was eliminated by rigging the spoilers 
symmetrically down below the wing surface. Thus, with flight airloads, 
the spoilers would float up no higher than flush with the wing surface. A 
system with no leakage is doubly advantageous since there will be no 
spoiler float either with flaps up. 

3.2.3 Full-Span Fowler Flaps 

Increased flexibility in wing-area design is achieved by the 
application of high-lift devices. On the ATLIT design, this flexibility 
is provided by 30% c full-span Fowler flaps. 

Two dimensional wind-tunnel development of this Fowler flap 
(reference 14) defined the slot gap and overlap for maximum lift in the 
6jr = 40° position (see table 3.3). This 2-D optimized gap and overlap 
is incorporated in the ATLIT design. Three-dimensional tunnel tests 
(reference 17), at a reduced Reynold's number using the optimum 2-D gap 
and overlap, failed to generate the expected maximum lift. A modified 
gap and overlap were defined which achieved the expected maximum lift 
on the 3-D model, but it is suspected that this anomaly is a Reynold's 
number effect rather than a 2-D to 3-D effect. 



49 



o 



Spoiler leak gap 
0,10%c(AVG)^/ 




Spoiler 



Spoiler leak gap 
0.13°^c(AVG) 



Spoiler leakage path seal 
(Foam rubber weatherstripping) 



64%c spar 



Figure 3.7.- ATLIT spoiler-leakage path with seals, 



TABLE 3.3 - FOWLER FLAP SLOT DIMENSIONS FOR WIND TUNNEL AND FLIGHT TESTING 



TESTS 


GAP 


OVERLAP 


REYNOLDS 
NUMBER 


2-D, REF. 4, OPTIMUM FOR b^ = 40° 

3-D, REF. 17. OPTIMUM FOR 6^ = 40° 

ATLIT TESTS, 6^ = 37.8° 
max 


2. 7%c 

2.2%c 

2.8+0.5% 


-0.7%c 

0.8%c 

0.7±0.4%c 


2. 2 X 10^ 
1.0x10^ 
2. 2 X 10^ 



The actual gap and overlap for the ATLIT has some span-wise 
variations due to construction tolerances. These are indicated below: 





Gap 


Overlap 


Average 


+2.8% c 


+0.7% c 


Maximum 


+3.35^ c 


+1.5% c 


Minimum 


+2.3% c 


+0.4% c 



The maximum flap deflection on ATLIT of tS-j ^^^ = 37.8° resulted 

T ) iTlaX 

from an NASA Langley Research Center, Quality Assurance Office, inspection 
requirement for certain clearances of the flap rollers in their tracks. 
The flap positions are illustrated in figure 3.8. 

3.2.4 GA(W)-1 Airfoil 

The development of the GA(W)-1, general aviation (Whitcomb), airfoil 
followed an iterative procedure of defining an airfoil shape, then 
evaluating its characteristics by the method of reference 6. The procedure 
required about 17 iterations to transform a Whitcomb 17-percent thick 
supercritical airfoil into one especially suited for a low-speed 
subsonic airplane (i.e., 200 knots top speed and RN = 6 X 10 ). The 
computer optimization of the airfoil emphasized low drag for lift 
coefficients ranging from cruise to climb and high maximum lift with 
docile stall characteristics. It was convenient, while not necessarily 
desirable, to retain the 17-percent thickness of the supercritical section. 
The final airfoil shape was evaluated in the wind tunnel (reference 5). 



52 



en 

Ca3 




Figure 3.8.- ATLIT airfoil and flap geometry. 



The GA(W)-1 17-percent thick airfoil is likely to be redesignated as 
the NASA-417 airfoil of the LS-1 (low-speed) family. This family, so far, 
includes airfoils which have design-lift coefficeints ranging from 0.2 to 
0.6 and thicknesses from 13 percent to 21 percent. The family is being 
expanded and several GA(W)-1 airfoils are being modified to have different 
camber and thickness distributions. 

For the ATLIT project, the 17-percent (GA{W)-1 section replaced the 
thinner airfoil of the standard wing, the 652-415 "laminar" section. 
Although this older 6-series airfoil was designed to have a laminar boundary 
layer over a large portion of the chord, the maintenance of conditions for 
laminar flow on a production light airplane is impractical. Thus, a 
comparison of section characteristics for the 6- and the GA-series airfoils 
with fixed transition is reasonable. 

The essential differences in the two airfoils are apparent in 

figures 3.9 and 3.10. In particular, for the GA(W)-1 section, i-r) is 

d max 

about 50-percent higher, C-. is almost 30-percent higher, and, 

max 

unfortunately, C. is about 6-percent higher and C is about 60-percent 


higher than for the older 6-series airfoil. 

The improvement in — with the GA(W)-1 section should contribute to 

d 

improvements in range and inclimb performance. 

The increase in section maximum lift permits greater flexibility in 

the design of a wing planform. Smaller wing areas may be designed while 

keeping the product, C, . S, constant. 

max 

A discussion of the effect of the higher C. appears in section 3.2.1. 





54 



+0.1 




-0.1 



en 



+0.1 r 




-0.1 L 



J I I L 



J L 



0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 



Figure 3.9.- Comparison of the GA(W)-1 and GS^-AIB airfoil shapes, 



Fixed Transition 
-,6 







Figure 3.10- Comparison of section characteristics of the GA(W)-1 and 65p-415 airfoil 



Increasing C was apparently regarded as unimportant, during the 
o 

GA(W)-1 development. It remains to be seen whether this assumption is 

valid. Increased GA{W)-1 C will cause increased trim drig. According 



to reference 23, longitudinal trim drag ranges from 1 percent to 10 percent 
of total airplane drag. For the cruise condition, trim drag will be in 
the lower part of this range, and during a climb, in the upper part. 
Therefore, the increased pitching moment of the 6A(W)-1 section is likely 
to produce an appreciable increase in total airplane drag over the whole 
operating range of the airplane. 

During the conceptual design of ATLIT, it was estimated that the 
weight of the wing could be reduced- about 10 percent. This reduced weight 
would result from the lower wing root bending moments for the tapered 
ATLIT wing and from the lower weight of materials for the smaller wing. 
In practice, this weight savings could be used to provide either increased 
single-engine rate of climb or increased useful load. On ATLIT, however, 
this potential was not realized. The design of the prototype-wing 
structure was done with little regard for weight. The resulting ATLIT 
wing is about 1.34 kN (300 lb) heavier than the standard wing. The 
conservative-design wing root bending moment for ATLIT is about 55-percent 
higher than for the standard Seneca. This moment is also about 
SO-percent higher than it need be for the design of such a wing 
to conventional limits. 

Early in the flight program, wing templates were made at eight 
span-wise locations on ATLIT to determine how well the actual wing sections 



7. Unpublished data in "Conceptual design of an Advanced Technology 
Light Twin;' 



57 



compare to the true GA(W)-1 shape. Figure 3.11 presents this comparison. 
In general, the ATLIT wing sections are representative of the GA(W)-1 
shape; however, the figure shows some discrepancies near the wing trailing 
edge. On the average, the ATLIT wing trailing edge is twice as thick as 
it was designed to be. The design trailing edge thickness varies from 
a maximum of 2.3% c to a minimum of 0.9% c with an average of 1.4% c. 
The ATLIT trailing edge thickness variations result from the difficulty 
in matching the position of the Fowler flap trailing edge with the wing 
trailing edge in the flap-nested configuration. The effect of these 
thickness variations will be most pronounced on section drag. Lift and 
pitching moment will be affected wery little. 

3.3 Supercritical Propellers 

o 

An analysis made on a modern light twin-engine airplane 
suggested that the efficiency of the installed propeller may be as much as 
5 percent to 15 percent (at cruise and climb, respectively) less than the 
efficiencies which larger, transport-category airplanes had been able to 
approach. A significant recommendation of this reference was that a 
propeller should be designed so that the twist and planform are optimized 
for operation in the flow field of the nacelle behind it. Such a 
propeller would improve climb and cruise performance. 

The propeller design was accompanied by a suggested design for modified 
engine nacelles which would achieve improvements in propulsive efficiency. 
Data of reference 22 indicate significant improvement in propulsive 
efficiency due to changes in the nacelle blockage effects behind a propeller. 

However, the design of modified engine nacelles for ATLIT was never completed. 

8. Correspondence from Mr. Howard Piper (Piper Aircraft Corporation) to 
Mr. John P. Reeder (NASA Langley Research Center). Subject: Advanced 
Technology Wings, Control Systems, and Propellers for General Aviation 
Twin Engine Aircraft, April 12, 1972. 



58 










Figure 3.11.- ATLIT wing templates, 



g 
Another recommendation was to consider both the older 

"traditional" sections (e.g., Clark Y, NACA 2412, or NACA 23012) 

and the newer supercritical sections for application to a new propeller 

design. The supercritical section was chosen to replace the airfoil of 

the standard propeller (approximated by the NACA 66-206). This choice 

was made in an effort to utilize the lift and drag characteristics of the 

supercritical sections to the best advantage in a propeller design. The 

design of the propellers was performed to provide an optimum combination 

of cruise and climb performance. Figure 3.12 illustrates the blade 

planform and airfoil used on the propeller. 

Additional anticipated advantages of the supercritical props include 

the more blunt blade leading edge and reduced blade weight. The blunt 

leading edge (in contrast to the sharp leading edge of the original blade 

section) will be less susceptible to damage from rocks and debris. The 

weight of the supercritical propellers is 27.8 N (12.5 lb) less per prop 

than the weight of the original propellers. A supercritical propeller 

weighs 291 N (65.5 lb) compared to 317 N (71.25 lb) for a standard 

Hartzell prop. 



9. Ibid. 

10. The propeller design was documented in the unpublished report 
"Conceptual Design of an Advanced Technology Light Twin". 



60 




Profile at maximum width 

17% t/c supercritical section 



Figure 3.12.- Supercritical propeller planform and cross-section. 



CHAPTER 4 

FLIGHT-TEST PROGRAM 

This chapter presents the objectives and. planning of the flight- 
test program and a description of the instrumentation system and 
flight envelope for the airplane. 

4.1 Program Objectives and Planning 

The flight evaluation of ATLIT has the objectives of determining; 

1. Stall speeds and characteristics 

2. Spoiler roll -control characteristics 

3. Cruise and single/multi-engine climb performance 

4. Longitudinal and lateral dynamic response characteristics 

5. Takeoff and landing distances 

6. Noise and performance characteristics of advanced technology 
propellers incorporating a supercritical airfoil. 

Items 1, 2, and 3 above are presented in this report. Items 
4, 5, and 6 are in progress at Langley and will be reported in 
proposed NASA technical publications. 

The flight- test program to complete items 1 through 6 above was 
estimated to require 100-flight hours in 50-calendar weeks. The 
purpose of the estimate was to provide data for support organizations 
at Langley to plan for such things as pilot man-hour and fuel require- 
ments. In retrospect, these estimates would have been quite accurate 
but for the impact of the tuft studies and instrumentation delays 



62 



63 
discussed in chapter 2.4. Revised estimates for the complete program 

increased the times to 150 -flight hours in 70-calendar weeks. 

An approximate breakdown is given below of the actual flight 

times required for completed tasks and estimated flight times required 

for planned tasks. The times listed under productive flight hours 

generated useful engineering data. The unproductive flight hours 

resulted in unusable data due to faulty equipment or procedures. 

About 30 percent of the total flight hours were unproductive. 



Total Productive Unproductive 



Task 

1. Airspeed and angle of attack 

calibrations 

2. Documentation of stall speeds 

and characteristics 

3. Spoiler roll-control evaluation 

4. Cruise and single-/multi-engine 

climb performance measurements 

5. Tuft studies for wing/body sep- 

aration cleanup 

6. Evaluation of longitudinal and 

lateral dynamic response 
characteristics 

7. Takeoff and landing distance 

measurements 

8. Measurement of noise and 

performance characteristics of 
the supercritical propellers 

TOTALS 

* estimated 



flight flight 
hours hours 



15 



6 


3 


16 


10 


26 


25 


21 


16 


(10*) 


- 


(15*) 


- 


(35*) 




(144*) 


61 



flight 
hours 



23 



63 



4.2 Flight Test Instrumentation 

The basic data recording system is described here. Special 

instrumentation and equipment required for airspeed calibrations are 

discussed in Appendix B. 

4.2.1 ATLIT Instrument Recording Package 

ATLIT was equipped to record on magnetic tape the 36 
flight parameters shown in table 4.1. The tape recorder had 
14 continuous, FM (frequency modulated) data tracks. One FM 
track was commutated to record as many as 28 channels of PAM 
(pulse amplitude modulated) data at a rate of ten samples per 
second. The approximate accuracies listed on the table are 
estimates of possible errors incurred between the in-flight 
measurement of a variable and the documentation of the variable 
on the ground in engineering units. Several possible sources of 
error are listed below. In general, the largest errors are 
caused by noise in the data system. The combined errors amount 
to about + 2% of full scale for each recorded parameter. This 
estimated maximum error represents three standard deviations. 
The sources of the errors are as follows: 

1. noise anywhere in the system (commutator, tape recorder, 
power sources) . 

2. sensor hysteresis, deadband, drift, and calibration 
accuracy (nonlinearity) . 

3. intermodulation errors during mixing of frequency signals 
for tape recording. 



64 



TABLE 4.1 - ATLIT INSTRUMENTATION PARAMETERS AND ACCURACIES 



u 

•I— l/l 

+-> Ol 

fO i — 

+-> JD 

I •>— 

+-> s- 

O fO 



on 
c I — 

O JD 
•1- fO 

+-) -r- 

o s- 



O) 


o 


u 


■* — 


s_ 


+-> 


o 


u 


U- 


OJ 




^^ 


1 — 


M- 


o 


<U 


s_ 


Q 


+-> 




c 


-o 


o 


c: 


C_3 


.fO 


s_ 




OJ 




5 


(jO 


o 


OJ 


Q_ 




1 


!q 


OJ 


tT3 


EZ 


■n" 




s- 


CT 


ra 


C 


=> 



Parameter 

Time 

Total air temperature 
Glfdeslope deviation 
Local izer 

Pressure altitude 
Static pressure 
Fine static pressure 
Airspeed 
Fine airspeed 
Vertical speed 



Range 

-18 to +38 (0 to 100) 
±0.7 
±2.5 

to 1524 (0 to 5000) 
to 103,421 (0 to 15) 
95,975 to 103,077 (13.92 to 14.96) 
to 200 
to 90 

*762 (+2500) 



Longitudinal acceleration 

Normal acceleration 

Lateral acceleration 

Roll rate 

Pitch rate 

Yaw rate 

Roll attitude 

Pitch attitude 

Yaw attitude 

Angle of attack 

Angle of sidesl ip 

Longitudinal wheel force 
Lateral wheel force 
Rudder pedal force 
Stabilator deflection 
Stabilator trim tab 

deflection 
Left spoiler deflection 
Right spoiler deflection 
Rudder 'deflection 
Rudder trim tab deflection 
Flap deflection 

Left eng.ine manifold 

pressure 
Right engine manifold 

pressure 
Left engine RPM 
Right engine RPM 
Left engine throttle 
position 



iO.SO 
to 4 
±0.50 
±100 
±30 
±30 
±180 
±35 
+ 30 
±20 
±20 

±133 (±30) 

±445 (±100) 

±667 (±150) 
-16 to +5 

±9 
-10 to 57 
-10 to 57 

±35 

±20 
to 40 



to 101,592 (0 to 30) 

to 101,542 (0 to 30) 
to 3000 
to 3000 

to 100 



Approximate 




Accuracy 


Units 


. 


Sec 


±0.6 (±1.0) 


°C (Op) 


±0.05. 


Deg 


±0.05 


Deg 


±30 (±100) 


M (Ft) 


±142 (i0.30) 


Pa (PSIA) 


±142 (i0.30) 


Pa (PSIA) 


±3.00 


Knots 


±1.8 


Knots 


±30 (±100) 


H/min (FPH) 


±0.01 


G 


±0.08 


G 


±0.01 


G 


±2.0 


Deg/Sec 


±0.6 


Deg/Sec 


±0.6 


Deg/Sec 


±3,6 


Deg 


±0.70 


Deg 


±0.60 


Deg 


±0.40 


Deg 


±0.40 


Deg 


±5 (±1.2) 


N (Lb) 


±18 (±4.0) 


N (Lb) 


±27 (±6.0) 


N (Lb) 


±0.4 


Deg 


±0.4 


Deg 


±1.2 


Deg 


±1.2 


Deg 


±1.4 


Deg 


±0.8 


Deg 


±0.8 


Deg 


±2031 (±0.6) 


Pa (In. Hg) 


±2031 (±0.6) 


Pa (In. Hg) 


±60 


RPM 


±60 


RPM 



±2 



Percent 



65 



4. excitation voltage error. 

5. analog to digital translation (ADTRAN) ground station 
nonlinearities. 

4.2.2 Nose Boom Installation 

A four parameter transducer instrument head is mounted 
on the ATLIT nose boom. The head senses: 

1. dynamic pressure (qc') for both cockpit panel and 
recorded airspeed data 

2. static pressure (p') for both cockpit panel and recorded 
altitude and vertical speed data. 

3. angle of attack (a') for recorded data. 

4. angle of sideslip (3') for both a cockpit indicator 
and recorded data. 

The noseboom, which is shown in figure 4.1, places the 
static pressure ports 1.7m (5.6 ft.) (approximately one maximum 
fuselage diameter) in front of the airplane nose. Data of references 
31 and 32 indicate that locating the static port at least one 
body diameter in front of the fuselage nose minimizes the position 
error. 

The instrument head used is typical of heads presently 
in use on NASA and other flight test aircraft. A detailed 
description of this head is contained in reference 33, and a 
summary of the pressure measuring characteristics of the head 
appears as table 4.2. The characteristics in the table apply 
at M = 0.6 (the minimum speed for which characteristics were 



66 



-3 



TOTAL PORT 

STATIC PORTS 
a-VANE 




(53 in. 

.1.448 m ^ 

(57 in.) 
L829ni H 

(72 in. 



2.032 m 
"(80 in.) 



Figure 4.1 ATLIT noseboom detail 



Table 4.2 Effects of Flow Angularity on the Pitot-Static Measurements (from ref. 28, M = 0.6) 



05 
00 





STATIC 
PRESSURE 
MEASUREMENT 
ERROR 


TOTAL 

PRESSURE 

MEASUREMENT 

ERROR 


Variation with angle of attack (3 = 0) 


+ 1 M 
for - 15° < Qt < + 35° 


- 0.2% 
for - 50 < a < + 20° 


Variation with angle of sideslip (a = 0) 


± 2.5"^ 
for -10° < 6 < + 10° 


- 0.2% 
for - 10° < 3 < 10° 


a = 0, B = 


+ 0.5'^ 


0% 



documented in the reference); however, for decreasing subsonic 
Mach numbers, the magnitudes of errors given in the table 
decrease. The static pressure errors shown are accounted for 
during the position -error calibration procedure of Appendix B. 

The effect of lag in the pitot-static system was measured, 
ensuring that the time-dependent behavior of the system would 
not have a significant effect on data recorded during nonsteady 
airplane maneuvers. The time histories for the pitot and static 
system responses appear as figures 4.2 and 4.3, respectively. 
The effect of lag in the pitot system was shown to be small in 
comparison to that for the static system. Therefore, only the 
effect of the static system lag is summarized as follows: 

Acoustic lag (time for a" pressure 

signal to travel through the 

static system) 0.033 sec. 

Pneumatic lag 0.056 sec. 

Static pressure transducer response 

time constant 0.094 sec. 



Total ATLIT static oressure system lag 0.183 sec. 

The small amount of lag in this system will result in 
less than 1.0% static pressure error (Ap/q' ) for nonsteady 
maneuvers in ATLIT which meet the following conditions: 

1. rate of change of airspeed less than about one knot 
per second at constant altitude or. 



69 



ATLIT pitot pressure system lag 



Input pressure transducer time constants 
Acoustic lag 



Total pressure 



o 



Output pressure transducer 
Time constant 



Input pressure transducer 
Time constant 



(pneumatic lag ^ 0) 
(tubing length ^ 5.0 m) 




Figure 4.2.- Time-dependent characteristics of the 
ATLIT pitot pressure measuring systems 



static pressure 



ATLIT static pressure system lag 
0.183 



Input Pressure Transducer Time Constant 



0.094 



^ 



Output pressure transducer 



'Time constant 0.15 



Output pressure transducer time constant 
Input pressure transducer time constant 

Pneumatic lag 

(Tubing length = 6 m) 



0.150 
■0.094 

0.056 sec 




63.2 percent of steady-state pressure 



pressure release 



Input Pressure Signal 



0.0 



0.1 



0.2 

Time t, sec 



0.3 



0.4 



0.5 



Figure 4.3 Time dependent characteristics of the ATLIT static 
pressure measuring system. 



2. rate of change of altitude less than 122 meters/mi n 
(400 FPM) at constant airspeed. 

4.3 ATLIT Flight Envelope 

The flight envelope for ATLIT is essentially the same as for the 
standard Seneca. The ATLIT was designed to at least meet FAR Part 23 
"normal" category limits. In order to adapt the new wing to the 
standard Seneca fuselage, it was most convenient for the main-wing 
bending member to cross through the fuselage at the same location as 
in the original airplane. For the symmetrically tapered ATLIT wing 
configuration, a line from root to tip through the carry-through spar 
contains the 50% chord points for all wing sections. As a result, 
as seen in figure 4.4, the quarter chord of the mean aerodynamic chord 
for ATLIT is slightly aft of that for the original airplane. 

The CG envelope for both airplanes is shown in figure 4.5. 
Empty weight for ATLIT is 13.26 kN (2,980 lbs, without instrumentation) 
This weight reflects an increase in ATLIT wing weight of about 1.33 kN 
(300 lbs). This increase is explained by the use of easily-designed, 
heavy, machined components and conservative assumptions in the wing 
design and construction. The ATLIT CG range in terms of fuselage 
stations was designed to be the same as for the original airplane. 
The limits relative to the mac for each airplane appear below (from 
reference 4): 



72 



-a 

CO 



Fuselage Stations 



Aft 
i 



E. - - Fully extended ATLIT flap 



137.828 



106.628-^ 



75.428 




Leading Edge of MAC - - Fus. Sta. 82.361 



Airplane Centerline 



ATLIT wing 



PA-34 wing 



Figure 4.4 Comparison of PA-34 and ATLIT planforms 



19- 



18- 



17- 



o 

■^ 16 



-M 

cn 

•r— 



15 



^ 14 

rd 
Q. 

^ 13-1 



12-1 

11 



4200 



3800- 



3400- 



o 



3000 



2600- 



Approximate ATLIT Empty Weight 
(with Instrumentation) 



Approximate ATLIT Empty Weight 
(without Instrumentation) 



Approximate PA-34 Seneca- I Empty Weight 



1 1 — I 1 1 1 1 1 1 1 1 1 1 I — 

80 82 84 86 88 90 92 94 
Fuselage Station 



T — I — I — I — I — I — 1 — I — I — I — r 



T — I — I 1 — r 



10 15 
PA-34 Wing MAC, % 



20 



25 



1 — I — I — 1—| — I — r 
-5 



T — I — I — I — I — I— I — r 



5 10 15 
ATLIT Wing MAC, % 



20 25 



Figure 4.5 Airplane CG Envelope for ATLIT and PA-34 mac's. 



74 



Fuselage % of % of 

Station PA-34 mac ATLIT mac 

Forward CG limit 4200 lbs 87.9 15.11 11.41 

Aft CG limits (all weights) 94.6 27.75 25,22 

The standard Seneca I is placarded for a maximum landing weight 
of 17.79 kN (4,000 lbs). Therefore, all weight added above 17.79 kN 
up to the gross weight of 18.68 kN (4,200 lbs) must be fuel. However, 
no such limitation applies to ATLIT. The fuel capacity of the modified 
wing is at least 568 1 (150 gal) compared to 379 1 (100 gal) for the 
original wing. Unfortunately, because of the large empty weight, 
the ATLIT full -fuel capacity can only be used with one person onboard. 

The permissible speed range for ATLIT was expanded slightly over 
that for the standard Seneca. For ATLIT, V^e = 211 knots versus 
189 knots for the PA-34 and VgQ = 51 knots versus 58 knots for the 
PA-34. The maximum flaps operating speed for the two airplanes remained 
the same at 109 knots. The use of full flap deflection (37.8°) on ATLIT 
for landing has been avoided because of the likelihood of the nose wheel 
touching down first. 



75 



CHAPTER 5 

FLIGHT-TEST RESULTS 

Flight-test results are presented in this chapter for ATLIT in 
essentially the configuration in which the airplane was delivered to 
Langley Research Center. The airplane characteristics reported on here 
include the following: 

1. Static pressure and angle-of-attack position-error 
calibrations. 

2. Stall characteristics. 

3. Roll characteristics, 

4. Cruise and Climb performance 

5. Pilot comments on stability and handling qualities. 

After the flight-test program began at Langley Research Center, 
several modifications were made to the airplane. These modifications are 
described as follows: 

1, Several devices {strakes, fillets, and vortex generators) 
were tried on the airplane to reattach a region of separated 
flow which was caused by interference effects at the 
wing/body juncture (see figure 5.1). 

2. Seals were added to the wing to reduce leakage of pressure 
through the gaps around the spoilers. (A description of the 
seals contained in chapter 3.2.2). 



76 



-a 

->3 



WING/ BODY FILLET 




DROOPED STRAKE 



rFUSELAGE AND WING 
VORTEX GENERATORS 




Figure 5.1.- Devices for wing/body flow attachment, 



3. The spoilers were rigged symmetrically down below the 
wing surface to reduce the drag penalty due to spoiler 
"float." (The rigging of the spoilers is described 

in chapter 3.2.2.) 

4. To reduce drag, the ATLIT wheel wells were fitted with 
balsawood blocks which formed wells similar in shape to 
those on the standard PA- 34. 

Flight data are not presented in this report for an airplane 
configuration with all of the above modifications. The effect of these 
modifications on most of the present test results is expected to be small. 
The possible exception is climb performance. These results will be 
reported in planned, future publications. 

Where appropriate, predictions of ATLIT characteristics are included 
in this chapter. Performance predictions were computed by three different 
methods; one rapid sizing procedure and two lifting line theory methods 
(references 29 and 30). The purpose in presenting the results of these 
different predictive techniques is to compare them to one another and to 
flight-test results. 

5.1 Position Error Calibrations for Static Pressure and Angle of Attack 

The position error calibrations presented here summarize the results 

of tests which are discussed in Appendix B. These calibration corrections 

have been made to all pressure and angle-of-attack data appearing in this 

report. 

The effects of both flap deflections and power changes on static 

pressure and angle-of-attack position errors are appreciable. The 

displacements of the calibration curves with changing flap position or 

n. Unpublished report: Loftin, L. K. (NASA Langley Research Center): 
Conceptual Design of Subsonic Aircraft, Chapter 6 - Estimation of 
the Size and Performance of Subsonic Aircraft, Feb. 1976. 



78 



power may be explained by the effects of these changes on circulation. 
According to the Kutta-Joukowski theorem of lift. 



^=pvr (5.1) 



or 



r -^ ^ ; (5.2) 

that is, wing circulation is proportional only to velocity. For a fixed 
airspeed, then, the position errors of a wing alone (as influenced by 
circulation) are constant. However, position errors for a three-dimensional 
airplane are also a function of flap deflection and power setting, as 
explained below. 

Vari^itiDn of Fowler-flap deflection will affect position error in 
two ways: 

1. As the Fowler flaps are deflected, the location of the lifting line 
(center of circulation) will move rearward. This effect changes 
the upwash conditions at the flow-sensor locations either 

ahead of the wing or ahead of the fuselage nose. The result 
is a shift in the calibration curve. 

2. Span-wise lift distribution is affected by changes in the 
fuselage and nacelle attitudes due to flap deflections at a 
given airspeed. These changes in lift distribution will also 
affect local circulation at the flow-sensor locations, causing 
a shift in the position-error curves. 



79 



The effects of power changes on position errors are explained by 
the influence of body attitudes and of propeller slipstream on span- 
wise lift distribution. As in (2) above, the result is a change in 
local circulation which affects the position-error calibration. 
5.1.1 Static Pressure Calibrations 

The corrections made to static-pressure measurements are given in 
figure 5.2 for all flap settings and for power-on and power-off at two 
flap positions. Little effect of landing gear position was detected 
during the calibrations; therefore, all calibration data are presented 
with gear up. Calibrated airspeed can be computed from the data on the 
figure using the equation 



■ / + AR. 



V^ = V/ /I + ^ . (5.3) 

c 



c c q 



Corrections to altitude (static pressure) data can be computed 
using the equation 

p = p' - (^ . q' (5.4) 

^'c 

where V' and q' are flight-measured values and ^-^ is from 
c c q ' 

c 
figure 5.2. 

5.1.2 Anqle-of -Attack Calibrations 

The corrections made to indicated angles-of-attack are given by 
the linear functions presented in table 5.1 and in fig,ure B-IO 
of Appendix B. 



80 




Indicated airspeed V^' , knots 



Figure 5.2.- Static pressure position-error calibrations, 



81 



00 

to 





TABLE 5.1- ANGLE-OF-ATTACK CALIBRATION EQUATIONS 




Flap 
Position 


Angle-of-Attack Correction 
from Indicated to True 


Approximate Limits of 


Linearity 


Power Off 


Power On 


6f= 0° 


a = 0.88 a' + 0.15 


a =11° 


a- 15° 


6^ = 10° 


a = 0.80 a' - 0:02 


a= 10° 


a= 12° 


6^ = 30° 


a = 0.82 ct' - 0.44 


a= 8° 


a = 11° 



5.2 Stan Speeds and Characteristics 

5.2.1 Predictions 

1 7 
The preliminary design estimates for maximum lift were based 

on section data for the GA(W)-1 airfoil and lift effectiveness 

data for a 30%c Fowler flap. Flaps up C. was predicted to be 1.8, 

max 

and flaps 20° C, (trimmed to forward eg for gross weight) was predicted 
max 

to be 2.6. A comparison of these predictions with flight- test measured 

values, presented below, shows fair agreement. 

After the optimum flap configuration for ATLIT was developed in 

two dimensional tests (reference 14), the final wing-spoiler- flap 

configuration was evaluated in three-dimensional tunnel tests (reference 17). 

A summary of flap effectiveness from 2-D and 3-D testing appears as 

figure 5.3. The figure shows a loss in flap effectiveness In going from 

2-D to 3-D configurations. For instance, at ^^ - 40 (no spoiler leakage), 

AC, is reduced by about 0.69 (or 30 percent) from the value for 
max 

AC-, . This loss can be explained by a combination of the following items: 
max 

1. The 3-D flap span is less than the airplane wing span. For 

ATLIT, b^-|ap/t>v^-jnq "" ^'^^ compared to 1.0 for the 2-D tunnel tests 

Thus, at 6^ = 40°, AC, is reduced by 

max 

AC, - 0.88 x AC. = 2.3 - 0.88 x 2.3 = 0.28, 
max max 

28 
or about q'^q = 40% of the total loss in maximum lift. 



12. Unpublished data: "Conceptual Design of an Advanced 
Technology Light Twin Airplane." 



83 



2 - 



INCREAAENT IN 

MAXIMUM LIFT 

COEFFICIENT, 

AC^, ACj 

max max 



00 








2-D WIND TUNNEL (REF.14) 
NO SPOILER LEAKAGE 
/-2-D WIND TUNNEL (REF.16) 
C5 0.23 %c SPOILER LEAK PATH 
3-D WIND TUNNEL (REF.17) 
NO SPOILER LEAKAGE 

3-D WIND TUNNEL (REF.17) 
0.3 %c SPOILER LEAK PATH 
ATLIT FLIGHT TEST 
0.23 %c SPOILER LEAK PATH 



10 20 30 40 
FLAP DEFLECTION 6^, deg 



Figure 5.3.- Comparisons of Fowler flap effectiveness from wind tunnel and flight tests. 



2. The 3-D wings on the tunnel model and on the airplane had 
3 degrees of twist and finite aspect ratio compared to no 
twist and infinite A for the 2-D tests. Using the method 
of reference 31 and the theoretical wing span loading for 

ATLIT (not accounting for the nacelles and fuselage) from 

13 
unpublished data , maximum lift was compared for the two- and 

three-dimensional cases. By such an analysis, it can be shown 

that C-j is reduced by about 0.1 compared to the C-i 

max max 

of the section near the mac of the wing. This reduction 

accounts for about 0.1/0.69 = 15% of the 3-D loss in 

maximum lift. 

3. Adding the fuselage and nacelles to the wing has an effect 
on maximum lift which is difficult to determine. On the one 
hand, addition of the bodies can be considered as providing 
additional (though small) lifting forces, thus increasing 
maximum lift. On the other hand, the interference effects of 
the bodies on the wing may reduce maximum lift. A method is 
given in reference 29 for estimating this lift loss. Treating 
each engine nacelle and the fuselage independently, the method 

''''''' \JviB\ax\ ^°-95- This ratio predicts a 

flaps up lift loss of 0.95 x 1.7 = 0.09 or ^^ = 13% of 

of the overall 3-D loss in maximum lift. 

13. Budish. 



85 



4. The 3-D flaps were constructed in span-wise segments compared 
to the one-piece construction for the 2-D tests. At each flap 
bracket location on the airplane, there is about 3 cm of open 
space (span wise) between the flap sections. No estimate is 
given here for the loss in lift due to these gaps. 

Although the data in figure 5.3 are presented for several different 
Reynold's numbers, direct comparison of flap effectiveness curves is 
still valid. This is true since the effect of Reynold's number and (in 
the case of flight data) eg location on flap effectiveness is negligible. 
5.2.2 Test Methods and Data Reduction 

The procedures for conducting stall tests outlined in reference 32 
were used as a guide for the ATLIT stall testing. Briefly, the power-off 
stalls were entered with throttles idled and with a rate of airspeed 
reduction not greater than 1 knot per second. 

In most cases. It was possible to define the stall point from flight 

records by a simultaneous occurrence of a "g"- and a pitch-break. 

Figure 5.4 illustrates such a case. Most stalls occurred at n < 1.0. In 

order to make a valid comparison between flight- and wind-tunnel lift data, 

airplane C. was computed based on the measured value of n at the 
max ^ 



(5.5) 



stall break, or 






max, A 


n 
. z 


• W 

• S 



86 



PITCH 
ATTITUDE 

20 




-10 



PITCH 
r BREAK 



INDICATED 
AIRSPEED 
V ', knots 



V, STALL 




VERTICAL 
SPEED 
W.FPM 




00 

-CI 



NORMAL 
ACCELERATION 



n^.g 



2.0 r 

1.0 
0.5 



I 



g-BREAK 



5 10 15 



ANGLE OF 

ATTACK 

a, deg 




STABILATOR 
DEFLECTION 

^STAB' ^'5 



STALL 



5 10 15 



.-v6 



STAB, STALL 




10 15 



TIME, sec 



Figure 5.4.- ATLIT stall time history (flaps up, approach power). 



The values of C, for several stalls in each configuration 
max, A 

were averaged. These averaged data appear in table 5.2. The maximum 
airplane lift data in the table were also corrected to the aft CG (25% mac) 
for both airplanes. This correction was made to minimize the trim-lift 
penalty in the data sO' that a more direct icpmparison can, be made. with wind- 
tunnel results. The correction was made using the following equation from 
reference 33: 



(C, ) " (C, ) + 
max, A max, A , 



C62 - CGj 
^ - CG 



J 



^^L J (5.6) 
max, A - ^ ' 



max r); 

where the quantities subscripted with a 1 represent flight-test values 
and those subscripted with a 2 represent the condition to which the data 
are being standardized. 

5.2.3 Results and Discussion 

As shown in table 5.2 , the highest lift coefficient attained by 

ATLIT was C, = 3.03 with 6^ = 37°. The corresponding stall speed 
max, A 

is V ^51 knots, which is 7 knots less than V for the standard 


Seneca I. With flaps up and spoiler leakage unsealed, C. =1.73. 

max, A 

The corresponding stall speed is V _ 68 knots, which is only 4 knots 

^1 " 

faster than the flaps-up stall speed of the standard Seneca I (which has 

35 percent more wing area than the ATLIT). The flaps-up maximum lift 

coefficient and the AC, due to flap deflection on ATLIT appear to 

max 



88 



TABLE 5.2.- COMPARISON OF ATLIT AND SENECA STALL SPEEDS AND MAXIMUM TRIMMED LIFT COEFFICIENTS. 



00 
CO 



CONFIGURATION 


SENECA 1 


ATLIT 




max, A 


max, A 




kts. (mph) 


kts, (mph) 


FLAPS 0° 


64 (7-4) 1.45 


_ 


SPOILER LEAKAGE SEALED 


- 


72 (83) 1.54 


SPOILER LEAKAGE UNSEALED 


- 


68 (78) 1.73 


FLAPS 10" 


- 


59 (68) 2.28 


FLAPS 20° 


- 


56 (65) 2.54 


FLAPS 30° 


_ 


53 (61) 2.81 


FLAPS 35° 


- 


52 (60) 2.87 


FLAPS 37° 


- 


51 (59) 3.03 


FLAPS 40° 


58 (67) 1.76 


- 



ALL DATA PRESENTED FOR e.g. = 25% M.A.C AND 
GROSS WEIGHT = 18 700 N (4200 lb) 



be the largest ever generated by an airfoil and a single-element flap of 
similar geometric configuration. This improved lifting capability of the 
ATLIT wing bears out the most useful quality of the new 6A(W) airfoils, 
that of increased flexibility in designing smaller wings. 

As illustrated in figure 5.3, wind tunnel testing (reference 17) 

predicted that sealing the spoiler leakage path would increase C. , 

max 

flaps up or down. However, as shown in table 5.2, the reverse occurred 

in flight. That is, when the spoiler leakage path was sealed with plastic 

tape, C. (flaps up) decreased by about -0.2. These results are 
max 

based on data averaged for less than a dozen stalls in each configuration, 

leakpath sealed and unsealed. It has been hypothesized that flow through 

the spoiler leak gaps provides a beneficial boundary- layer control effect 

to postpone flow separation in flight. Another possible explanation is 

from the effect of small spoiler deflections on wing lift. As discussed 

in the spoiler description of chapter 3.2.2, separated flow behind a 

small spoiler deflection can reattach before reaching the wing trailing 

edge. The net effect is increased camber and increased lift. The stall 

tests discussed here were done with the spoilers floating symmetrically 

up above the wing surface by about 2 to 3 cm. This small float angle 

could be responsible for the increase in C, with the spoiler 

max, A 

leakpath unsealed. This phenomenon will be investigated in planned- 
full-scale wind tunnel testing with ATLIT. 



90 



A direct comparison between 3-D wind-tunnel data (reference 17) and 
flight data on flap effectiveness is made in figure 5.3. The figure 
shows that the ATLIT flap effectiveness (with a 0.23% c spoiler leak gap) 
agrees closely with that measured on the reflection plane wind-tunnel 
model (which had a 0.36% c spoiler leak gap).. The figure also shows 
consistently increasing lift increments with increasing flap deflections. 

Qualitative comments on stall characteristics appear in chapter 5.5. 



91 



5.3 Spoiler Roning Characteristics* 
5.3.1 Spoiler System Development 

The following discussion will cover many of the points discussed 
in chapter 3.2.2, but with the emphasis on the historical development of 
the ATLIT roll -control spoiler system. 

When construction of the ATLIT wing was started, only the span, chord, 
and location of the spoilers at the 70-percent chord line had been decided, 
The wing was therefore under construction with a hole provided behind 
the 70-percent chord line for the spoilers. The Mitsubishi and Redhawk 
spoiler configurations were being considered for use on ATLIT. However, 
at that point, W. H. Wentz, Jr. and H. L. Crane, among others, became 
concerned about the need to minimize the nonlinear spoiler characteristics 
which would be induced by the large full-span Fowler flaps. Therefore, a 
two-dimensional investigation of several proposed spoiler configurations 
was made in the Wichita State University wind tunnel. Unfortunately, no 
hinge-moment data were obtained. Design details for the ATLIT spoiler 
system, such as spoiler cross section, spoiler vent path, and spoiler 
leading edge gap, were then based largely on the WSU tunnel data. One 
feature, the leading edge gap, was adapted from the Mitsubishi, MU-2 
spoiler design and, therefore, may be covered by the MU-2 patent. The 
aforementioned two-dimensional WSU data were reported in reference 15. 

It should be noted that Wenzinger and Rogallo, in NACA TR 706 
(reference 37) examined spoiler configurations similar to the one 
adopted for ATLIT (except for the absence of leading edge gap). 



*The contributions of Mr. Harold L. Crane (NASA-LaRC) in preparing 
materials for this chapter are gratefully acknowledged. 



92 



Conclusion 1. from TR 706 was as follows: 

"Spoilers alone were found to be generally unsuitable for 
lateral control on wings with full-span split or slotted 
flaps because of excessive lag and because of ineffective- 
ness at small spoiler projections. The characteristics 
were improved as the location of the device was moved 
toward the trailing edge of the wing. Spoilers alone may 
give acceptable control for some types of airplane if 
they are located sufficiently near the wing trailing edge," 

Therefore, the LaRC monitors also arranged that before ATLIT flew, 
Paulson (reference 16) would measure the lateral control characteristics 
of the ATLIT spoiler configuration on an available rectangular wing with 
an aspect ratio of 9 and a GA(W)-1 airfoil section. The nonlinearities 
of flap-down spoiler effectiveness with increasing spoiler deflection 
measured by Paulson were found to be similar to the Wentz two-dimensional 
results. 

A second NASA contractor report (reference 17) by Wentz, et. al . , 
which presents spoiler characteristics including hinge moments on a 
one-fourth scale model of the ATLIT wing panel is now being printed. This 
paper confirmed that, as Calhoun* had estimated, the spoilers would have 
a positive (i.e., self deflecting) hinge moment over much of the deflection 
range. This undesirable characteristic persists to larger deflections 
flaps down than flaps up. 



♦John T. Calhoun of Robertson Aircraft Corporation, the designer of the 
ATLIT spoiler roll -control system. 



93 



As a further precautionary step, the newer Princeton variable- stability 
Navion» which was developed using NASA LaRC funds, was configured to 
evaluate pilot tolerance to various degrees of nonlinear roll control. The 
results of this in-flight simulation were reported by Ellis, et. al . , in 
reference 19. 

In summary, then, the Calhoun design estimates as well as the several 
wind-tunnel investigations showed that the spoilers would provide powerful 
lateral control with undesirable variations in the rolling effectiveness 
derivative 3C,/96 between small, medium, and large spoiler deflections. 

However* the tiiro -dimensional tests at WSU made it possible to select spoiler 
vent-path geometry to both eliminate reversals of spoiler effectiveness 
at small deflections, and also to always provide at least some spoiling 
of lift, which resulted in some pro-rolT rolling moment at small deflections. 
Also, the flight simulation at Princeton University showed that the very 
experienced NASA research pilots could cope with extreme nonlinearities 
of lateral control effectiveness during approach and landing In strong, 
gusty crosswinds. Apparently they were able to concentrate entirely, or 
solely, on airplane response and almost instinctively move the wheel as 
required to get the desired lateral response. 

In the rest of this chapter, plotted examples of the estimated and 
measured spoiler stability and control characteristics from the 
aforementioned sources will be shown and discussed in more detail- 

5.3.2 Methods and Data Reduction 

All spoiler roll -control data presented here were generated (rudder 
locked) by abrupt spoiler inputs with an adjustable wheel travel limiting 



94 



chain attached to the pilot's control wheel. These abrupt spoiler inputs 
(in practice) were ramp inputs of about 0.2-second duration. The 
combinations of configurations and airspeeds tested are given in table 5.3. 

Two types of rolling maneuvers were used for the roll tests. At 
first, rolls were initiated from a bank angle of 30 degress and were 
allowed to continue to a bank of 30 degrees in the opposite direction. 
This conventional bank-to-bank type of maneuver proved to be undesirable 
for gathering spoiler roll data. In the second type of rolling maneuver, 
the roll was initiated from level flight and was allowed to continue to 
about 45 degrees of bank. This second type of maneuver was favored over 
the first for several reasons. With a sample rate of 10 per second on 
all recorded flight parameters (see chapter 4.2), the pertinent data for 
analyzing roll performance are gathered during the first second or so 
of the maneuvers; thus, no real need exists for the long duration of 
roll rate provided in the bank-to-bank type of maneuver. More important 
is the fact that spoiler produced rolling moments are sensitive to 
angle of attack; therefore, angles of attack must be consistent and 
accurately defined during the maneuver. Beginning the roll from a 
wings-level attitude simplifies data reduction since it is then not 
necessary to account for a varying load-factor- induced angle-of-attack 
(varying as bank angle varies) encountered during the bank-to-bank type 
of maneuver. In addition to these reasons for favoring the maneuver 
with a wings-level initial condition, the rolling maneuver from wings 
level is easier to perform from a piloting standpoint. Establishing a 
wings-level initial condition for the rolling maneuver requires less 



95 



TABLE 5.3- CONFIGURATIONS AND AIRSPEEDS FOR SPO.ILER R Q_LL TESTS 
Flap setting, deg Calibrated Airspeeds, V", knots 

79. 100, 123, 148 

10 64, 79, 95 

20 64, 78, 95 

30 59, 69, 76, 95 

37 55, 64, 78, 95 

Note: The data are largely for rolls to the right with spoiler 
deflections up to about 55°, i.e., 6^ = 50° or 



max 



> 

Ah/c = .05 (with c referenced from the w.inig chord at -the 

midspan of the spoilers). 



96 



effort and is less time consuming than establishing a banked-initial 

condition. 

In order to present data in the form of rolling moment coefficients, 

the roll damping coefficient, C-, , was first required for ATLIT in all 

P 

configurations of interest. The C, estimation procedure and results 

P 

are presented in appendix C. Then, C, can be determined simply by 

C. = -C, • {^) . (5.7) 

p 

The helix angles (Srr) for equation (5.7) are based on flight-derived 

values for roll rate (p) and true airspeed (V). 

5.3.3 Flight-Test Results 

Because the wing designers allowed inadequate room for the spoiler 
and flap systems, the ATLIT spoiler actuating system had to be crowded 
into a space which was inadequate both chordwise and vertically. Largely 
as a result of this crowding, there is an excessive breakout force of 
about 44 N (10 Ibf) in the lateral -control system. Therefore, the reader 
should keep in mind that the ATLIT spoiler roll-control system, although 
it is wery powerful, represents an example of a far-from-optimum spoiler 
system. 

As with any cable-driven system, the amount of stretch in the system 
(i.e., the reduction in maximum deflection of the control surface) is 
roughly proportional to the dynamic pressure or to a squared function of the 
airspeed. Figure 5.5 shows that the maximum stretch was about 30 percent 
which, because the authority of the spoilers was so great, was considered 
to be readily acceptable. 



97 



SPOILER 

DEFLECTION 

WITH FLIGHT 

AIRLOADS 6 ,deg 






60 



50 



40 



LINE FOR NO SYSTEM STRETCH: V =0 




30 



20 - 



SPOILER DEFLECTION WITH NO AIRLOAD 6^, deg 



Figure 5,5- Variation of ATLIT spoiler system stretch with airspeed 
(flaps up, spoiler leakpath sealed) 



As of this date, the airplane has been flown with the spoilers 
rigged several different ways. The characteristics are very strongly 
influenced by the spoiler rigging. To be on the safe side, the early 
flights at Piper were made with the spoilers rigged up 8 degrees to be 
certain that there was no dead band in effectiveness. The pilot comments 
about the lateral control with the spoilers rigged 8 degrees were ^^Qvy 
favorable. Piper flew about 20 hours including the ferry flight to 
Langley Research Center with the spoilers still rigged up 8 degrees; they 
had to make a crosswind takeoff and landing enroute and did so with no 
problems. While the airplane was being instrumented at Langley (LaRC), 
the spoilers were rerigged to be flush with the wheel centered. All the 
flying done by LaRC pilots was with the spoilers either rigged flush or, 
to provide a possible drag improvement, with the spoilers rigged 
symmetrically down 2 degrees. All of the roll -control data presented in 
this report were obtained with the spoilers rigged flush with the top of 
the wing with no air loads. However, some of the unfavorable pilot 
comments given in chapter 5.5 were obtained with the spoilers rigged 
down slightly for possible drag improvement. This rigging was discussed 
in chapter 3.2.2. 

Figure 5.6 shows two example time histories of the roll response to 
spoilers of the ATLIT airplane. These rolls were made at 78 knots with 
the flaps down 30 degrees. The spoiler deflections for the two rolls 
were about 15 degrees for the small input maneuver and almost 50 degrees 
for the large deflection roll. It can be seen that it took the pilot 
about three- tenths of a second to deflect the spoilers. There was a 



99 



Yaw rate 



Roll Rate 
p dea. op 

sec *~^ 




Bank Angle 
*. deg 



150 
100 

Lateral wheel ^0 ^ 
force F„ . N 

"y -50 [- 
-100 t- 



lb 
-25 



Spoiler deflection 

5,. deg 




-10 



I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I I ll 1 1 I 1 I I 1 I I I I I 1 IIJ I I N I 1 I I I I 1 1 I LlJ I I 1 1 1 1 I I 
12 3 4 5 6 7 

Time t, sec 

(a) Large spoiler deflection 
Figure 5.6- Roll response time histories 



100 



10 
Yaw Rate 



■10 I — 



Roll Rate 

P ^ 
sec 



25 



-25 



r 



Bank Angle - 
*. deg 




100 p- 

Lateral wheel IT" 
force FN h— 
y -50 — 

-100 t- 



25 

lb 

-25 



20 p- 



Spoiler deflection 
6,. deg t 




M I II I I I n I I p I I I I I I I I I I I I I 1 I I I 1 I I t I I I I I I I I I I I I I I I I I I I I M I I I 1 1 I I I 

12 3 4 5 6 7 

Time t, sec 

(b) Small spoiler deflection 
Figure 5.6- Concluded. 



101 



breakout force of about 40 Newtons in each case. For the smaller input 
maneuver, the roll -control force went to zero in the steady part of the 
roll. In each case, the bank angle varied smoothly from a left bank to a 
right bank. These rolls were started from 25-degrees left bank and, in the 
case with the small spoiler deflection, continued to about a 30-degree bank 
to the right. The maneuver with the large spoiler input was continued to 
a right-bank angle of about 70 degrees. In each case, it took about 
1 second for the roll rate to reach its maximum value. At the top of 
the figures, the yaw rate shows no sign of adverse yaw for either roll. 
As explained elsewhere in the paper, the raw data shown here cannot be 
used to determine the phasing of the various parameters. Reduction of the 
data introduces an erroneous phase difference of as much as two-tenths 
of a second between different parameters. The only undesirable 
characteristics shown by these time histories appear on the force trace 
for the smaller deflection roll. A large breakout force of approximately 
40 Newtons showj/!s clearly and the time history also shows that the force 
goes to zero during the steady part, of the maneuver. Except for this 
undesirable characteristic, everything else shown on these two plots 
indicates characteristics of a satisfactory lateral -control system for 
rolling maneuvers. 

Figure 5.7, a five-part figure on 5 pages, shows the measured 
maximum rolling velocities as a function of spoiler deflection for flap 
deflections from to 37 degrees and several airspeeds as were listed 
in table 5.3. Most of these data are from rolls to the right. However, 



102 



70- 



60 



50 



DEG/SEC 



40 



30 



20 



10 



ffliimiiiii 



Flagged symbols indicate rolls to the left 



11 




^s,DEG 



.01 



.02 



03 .04 

Ah/c 



05 



.06 



(a) Flaps O" 
Figure 5.7- Rolling velocities as a function of spoiler deflection 



103 



^Flagged symbols indicate rolls to the left 



DEG/SEC 




01 



02 



20 


30 40 




<Ss,DEG 


1 


I 


.03 


.04 
Ah/c 




(b) Flaps 10° 


Figure 


5.7- Continued. 



05 



.06 



104 




01 



02 



.03 .04 
Ah/c 



.05 



.06 



(c) Flaps 20° 
Figure 5.7- continued 



105 



DEG/SEC 




01 



,02 



__J i_ 

.03 .04 
Ah/c 



05 



,06 



(d) Flaps 30° 
Figure 5.7- continued 



106 




01 



.02 



— I 1_ 

.03 .04 
Ah/c 



.05 



.06 



(e) Flaps 37° 
Figure 5.7- Concluded, 



107 



data from a few left rolls are shown and can be Identified by the flagged 
symbols. Figure 5.8 shows the same roll data converted to the maximum 
roll helix angles^ SS-, attainable with the varying spoiler deflections. 

The original Gil ruth standard of reference 38 was that for acceptable 
roll -control power pb/2V should be at least 0.07. This standard still 
remains a good one for most general aviation airplanes. It is apparent 
from figure 5.8 that the roll authority of the ATLIT spoilers 

(|&) = 0.10 to 0.15) far exceeded the Gil ruth standard for large flap 

deflections. A spoiler will obviously become less effective as the portion 
of the wing behind the spoiler becomes less heavily loaded. Therefore, 
it was to be expected that P^/ZV would be lower for a flap deflection 

of 10 degrees than for larger flap deflections. However, the Gil ruth 

standard was still exceeded with ^ = 0.075 for 6^ = 10 . With 

flaps up, the ATLIT pb/2V„^„ response was about 0.08 at cruising speeds, 

max 

but was slightly below the standard with a pb/2V_^ of 0.055 at a speed 

max 

of 80 knots for 6 - 50 . However, It should be noted that an actual 

spoiler deflection of" 60 degrees could probably be used to Increase the 

attainable pb/2V„,„ at 80 knots to at least 0.065. 
"^ max 

Figure 5.9 presents the roll -control authority of the ATLIT spoiler 

as a function of airspeed for 40 degrees of spoiler deflection. The 

figure shows that even though 40 degrees is less than 80 percent of the 



108 



£b 

2V 
RAD 





■■:::;: 


:::: : ;:::;;;:;-;: 


" : 


FHT 


...... 


1 H7 W ^W 1-1 )m4 Hl-t HH 1 fH H H hm4 If 


FT+H 


-T4-f ff+H4 


H4ff 


+H4 441-1 




::::::■ 


mm|m4uiM4wmwmmHttuwiMJttWH 


1-OTfctt 


::;:::;;:::;::::::■ 




::"::::: 


0148 KTS 
Q124 
O 95 
A 79 


Br 




::::::;::::::; J i;::};!:!! :.T:::: :::::::::: 


::::::::;:;:?::■■ 


•14 


:.,:;;;;;:::::;i::;;:;;::ii:::ii.::.i::^:;;:;;;;;:;;;; 




n III 


;;: 


fflrtrmmttilitW WMW [If flHSt tH 1 WHIWIUMIH 




:::::::: 


i; '■':- 


:: ;;;|:; i;:.;^ :;:;:■;::: :;:■• : :;. ■: : :.::::::::::.::::::_;::::::: 






''H 


- : . /:-:;; J: - - - - ■;-:... | ::;:: :f ,.:;.;:-::::::.:::■ ■ 


.12 


itnni 


M^^ 


";■::: : - : : " - " ' ' ' ■ " ':: '■.:-■ ':' 'W \'\\]:'.:\\:\ .]:.:'' - 


;-■■;;:; 


ll -' 


■■'■'": ':'■':' ..\ .;.::::.;;,■ ■■■ ■ .■ ::■ :\:.\'\'-i\'\ ::[■[[ 




: :: :: 




::■.]::'. 


■i ■'■■■ i-' ' ■ -h- ■ '-^ ' ■ - '■ ■if;;^;:;::;-.;-- 




■ It 


|..:]:.:::::^:: 


:::::■-:. 


iiifHiii fi liiiil}[i m m rfiilt InffllfHili 


.10 


: . . _ : 1 




-;l : . ■■": : ; :■■ ■.::.: : |l . " If. : ■:'!":: : ;;;:t''F''"i 


v;'^-; 


■■■-ft ' i 


{.' : :-\ :■ ■■]{'- ■■- -f ■ 


4m w^ 




M - 


h^ + :\^-. }{:;:}.■ J 


:::::;:::::;::::;■::;•::■:• ■:;::::::::::::; 1 




V ■■■:;::. : J " - v ;| i: | ; i I: ' -'F-" 'l '^ ;■:.; J ■:}[ |, ,■'"': T :|I:.-|::r^: 


.08 


^^■:-: : :^:::-| J -- ::::_...■.:■:: ■■:::? g !. ^ ,,,^ •,■;::/: j" ;TT::: 


''■■■■'' t" -:;;,::..:, tj ; tI^ ' f ■ ''?m|^'''''' i ■■''^■|r^ 




-|- ■■:...■:_■_._:-:_■ r::::..: . { : . ,,!i: |;,';:| :: ^ -:; .. ^^'^^ J .. 




.. ::-: 


] ■■' I::.::::: ijz 


::;;: ::^^::::::;::;;-^::::: M,}lf.^^,,:....m 


: ::::::;::;.■. 




i £■-■ ill 


.06 


.;:. : 1 f.: :" | ■:-_; |: :: .:: : |, J ^: ::. :; i^' ' J ^ | :J;f '' - : J]f ."[:: -:'^ i:":;;;!-; 


: -1" - ■ ]_■■::": "|l; 


. . .. . T 


;::.4.:;»:.,:.'i:-:,::^^:^: ::/:;..,■■■■■■ 


^': H -jf;;:-' 




I-- 




J||f(j| [t| Rfm mfftefljlf^^ 


"■ ." ■'[ .'"1 




■ - - - -■ ■ ■ ■ :;i:;:::;;, ;::::: ■:;:?■ :::■■...:;;■::.:- :. . 


f '1^ . M- ":i^- 


0^ 


|i ||{i [ ffiyijijiffliiffl [H 


. . ; ■ ■ -T - "" r " 




1 III k| Wjffli [SfflWi|MffBM (flmilfit fi P 


: ■ .: ;i 




, _ - ,i^: ;. 1 : .'^ihi! :;;;■'= .. : ■ : : . j 


^■rtfttr^" 




^ ^i 1 .l' ;|' |m': ,; |.: |;;;i::':|: : 1^^^ - - IS "^ i ^U t 


i ■ttr '■;! — ■■ 


n? 


i.:|il. ,; '.,%''' \.'..:\\\' '-' -;-■:--:::-■::-]._, - f - -■ J,; 


t 1 






: T __ 1 -,:':;;' ' J!''; f: ■■!|;t-T" ;■"; ;" ::- :■■■■■;+■■■:;■ -4- 




1- :||!^ l!li V. 






|-H. 


: :' '- : 4 - : - " -' : "" . : 






■ :i r : 




|^^ffl|l^|||WM 




■:l^ - - r'-^-':.: W' j_ "t. £ t 




!{:-; 


f- 


mw 




i'^f .. ■ I :■[[ :_| ;i 


L::.:J:J 


.:.:::^:;:;"^. ■:;:;. I' J 


.:.; 


-■■■"A f ' 

- N f 


kMm\::^ 



10 



20 



30 



40 



50 



60 



,01 



.02 



.03 .04 
Ah/c 



.05 



.06 



(a) Flaps 0° 
Figure 5.8- Roll helix angles for varying spoiler inputs 



109 




01 



02 



.03 



.04 



.05 



.06 



Ah/c 



(b) Flaps 10° 
Figure 5.8- continued 



110 



2b 

2V 

RAD 




.01 



02 



.03 .04 
Ah/c 



.05 



06 



(c) Flaps 20 
Figure 5.8- Continued 



111 




20 



30 40 
^s,DEG 



50 



60 



01 



,02 



.03 .04 
Ah/c 



.05 



06 



(d) Flaps 30^ 
Figure 5.8- continued 



112 



RAD 




fig.DEG 



X 



.01 .02 



,03 . .04 
Ah/c 



.05 



,06 



(e) Flaps 37° 
Figure 5.8- Concluded. 



113 



ht^ 



.14 



.12 



.10 



HELIX 
ANGLE 



2V 



, RAD 



.08 



.06 



\ 



5. = 37' 
'"^ f 



5f = 30*^ 



5^ = 20' 



VTTTT 



^^ = 10° 



Reference 38.- lateral handling 
qualities specification 



6f= 0^ 



04 




L 



J « « ' 



J -I f I \ 



60 



80 



100 



120 



140 



160 



V^, knots 



Figure 5.9- The effect of flaps on ATLIT rolling performance for 40^ of 
spoiler deflection (6^ = 0°, 10°, 20°. 30°, 37°). 



net spoiler deflection available on this airplane<^after allowing for 
cable stretch, the rolling performance is generally well above the 
desired pb/2V =0.07 level. It can be readily determined from this 
figure that the roll -control authority is strong, or at least adequate, 
with the exception of the clean configuration near the stalling speed. 

In order to be able to determine values of C/ , the rolling moment 

coefficient, C, , the roll damping coefficient (flaps up) was calculated 
P 

using the methods discussed in appendix C. Values of C, for flap-down 

P 

configurations were estimated based on the projected planform area for 

each flap position. The aspect ratio and taper ratio were adjusted 

accordingly for the calculation. The calculated variations of C, 

P 

with C. for the flaps-up case and for 6^ = 10° and 30 are presented 

in figure C.3 of appendix C. The ranges of the estimated values of C, 

P 

over the appropriate ranges of C, were approximately 0.5; + 0.05, with 

flaps tip and approximately 0.6; + 0.04, -0.06 with 6^ = 30°. 

Figure 5.10 shows the variation of C-. ' for ATLIT with spoiler 

deflection for three flap deflections and several speeds. It should be 
noted that the spoiler deflection is given both in degrees of arc and in 
terms of exposed height of the spoiler above the wing (using the wing chord 
at the spoiler mid-span as a reference length). Of these measures, spoiler 
projection in percent chord is a much more significant indicator of 
spoiler effectiveness. These data were determined using faired values 



115 



-.10 



c 



Flight-test data 



' Wind-tunnel data (ref. 17) 



-.05 



a = 8° 




V(, = 124 kts (a = 1.9°) 
V^, = 95 (5.0°) 

V^ = 79 (8.4°) 



0.02 



0.06 



0.04 
Ah/c 
(a) Fowler flaps nested 
Figure 5.10 ATLIT spoiler produced rolling moments 



116 



-.lOr 



-.05- 




Vj, = 95 kts (ct = 0.3°) 
V^ = 80 (4.3°) 



0.06 



Ah/c 
(b) Fowler flaps 10° 

Figure 5.10 continued 



117 



-.i5r 



Flight- test data 
Wind-tunnel data (ref. 17) 



-AO - 



C 



-.05 



(c) Fowler flaps 30° 
Figure 5.10- Concluded. 



(, = 68 kts {a = 0.4°) 
V^ = 59 (2.7°) 




60 



H8 



of pb/2V. As would be expected, the C ' curves are quite similar to 

the roll-rate plots of figure 5.7 or the pb/2V plots of figure 5.8. The 
designer, who plans to use spoilers, should be on the lookout for unusual 
trends near the stall such as occurred with cS^r = 30 at an airspeed of 

53 knots. In that case, the value of SC-, '/BfS was smaller than usual 

at small deflections, but it became larger than usual at medium-sized 

deflections. It should be noted that nonlinearities of the magnitude 

shown here could cause the pilot to overcontrol . On ATLIT, this tendency 

is compounded by the existence of a large wheel -breakout force- These 

ATLIT deficiencies are discussed further by the research pilots in 

chapter 5.5. Discrepancies between wind-tunnel and flight values of 

C ^ in figure 5.10 are most likely due to inaccurate estimations of C, . 

P 

Figure 5.11 summarizes, in terms of numerical pilot ratings, the 
conclusions from the Princeton in-flight simulation of nonlinear roll- 
control effectiveness of reference 19. However, it should be noted that 
the variable stability Navion used for these tests had a very smooth 
lateral-force characteristic with negligible friction or breakout force. 

On the favorable side, it should be noted that for ATLIT the 
characteristic variation of C ' with 6 was never as nonlinear as the 

one-fourth scale wind-tunnel tests indicated. (Examples of the wind-tunnel 
data of reference 17 are shown on figure 5.10.) 



119 



(U 

E 
(U 

o 

Id 
a. 



max 



-o 







Conventional 100% 18% 
Navion 




max 



o 



Low 

Inertia 

Case 

Pilot 

Ratings 



High 

Inertia 

Case 

Pilot 

Ratings 



Low 

Inertia 

High 

Crosswind 

Ratings 



6^^ - NAVION WHEEL DISPLACEMENT (DEG) 



8 - 9 



3J5 - 4 



4% - 5 



Figure 5.11- Pilot rating for nonlinear roll-response variations (reference 19). 



5.4 Cruise and Climb Perfonnance 

5.4.1.1 Method A: Performance Predictions* 

The method used to make these predici tons of airplane performance is 

14 
a rapid sizing procedure. A brief description of some assumptions 

and procedures used in the method is presented here. 

The procedure makes use of correlations of characteristics for 140 
different present and past aircraft. These aircraft were classed into 
different groups according to aerodynamic sophistication. ATLIT was 
classed with a group of aircraft having internally braced wings and 
retractable landing gear. A comparison of ATLIT to other airplanes in 
this class showed that the ATLIT characteristics (zero lift drag, aspect 
ratio, and wing loading, in particular) fell near the extremes of the 
ranges of these characteristics for this airplane class. The effects of 
this result are not accounted for in the performance predictions which 
are given in table 5.4. 

Maximum level-flight speed can be estimated by rewriting the equation 
for power required for level flight 

550Pn = Cp fi SV'^ (5.8) 



as 



V = 77 3 y^^ n ' (5.9) 

^ ''^ W/P ^C^ 



*The contributions of Mr. Laurence K. Loftin, NASA-LaRC, in the 
preparation of materials for this chapter are gratefully acknowledged. 

14. Loftin. 



121 



TABLE 5.4- COMPARISON OF PIPER SENECA I PERFORMANCE WITH PREDICTIONS FOR ATLIT.* 



DO 
CO 



Performance Characteristic 

{Sea-level values unless 
Indicated otherwise.) 


(1) 
Piper 
Seneca 
Handbook 
Values 


(2) 
Method A 
(Loftin) 
Using Piper- 
Seneca 
Geometry 


(3) 
Method A 
{Loftin) 
Using ATLIT 
Geometry 


(4) 
Method B 
(Ref. 30) 
Using 

Light-Twin 
Wind-Tunnel 
Polar 
(Eqn. 5.17) 


(5) 
Method B 
(Ref. 30) 
Using NCSU 
ATLIT Drag 
Polar 
(Eqn. 5.16) 


(6) 
Method C 
(Ref. 29) 
Using 
Seneca 
Geometry 
{Fig. 5.16) 


(7) 
Method C 
(Ref. 29) 
Using 

ATLIT Wing 
Geometry 
(Fig. 5.16) 


Max. level-flight speed, 
knots, (mph) 


170 (195) 


182 (210) 


200 (230) 


178 (205) 


177 (204) 


154 (189) 


173 (199) 


Max. rate of climb, 
m/min (ft/min) 


415 (1360) 


475 (1560) 


487 (1598) 


496 (1626) 


518 (1701) 


470 (1570) 


517 (169$! 


Single-engine rate of 
climb, ra/min {ft/min) 


58 (190) 


140 (460) 


152 (500) 


HI** (364) 


143** (469) 


143 (470)- 


163 (530) 


Best rate of climb speed, 
knots, (mph) 


91 (105) 


71 (82) 


70 (80) 


100 (115) 


96 (111) 


94 (108) 


94 (108) 


Best single-engine rate of 
climb speed, knots (mph) 


91 (105) 


71 (82) 


70 (80) 


92** (106) 


85** (98) 


94 (108) 


94 (108) 


Best range speed, 
knots, mph 


95 (109) 


94 (108) 


92 (105) 


99 (114) 


91 (105) 


95 (109) 


92 (106) 


Service ceiling, 
m (ft) HSL 


5486 (18,000) 


6103 (20,025) 


6422 (21,070) 


5999 (19,681) 


6866 (22,525) 


- 


- 


Single-engine service 
ceiling, m (ft) HSL 


1113 (3650) 


2551 (8370) 


Z883 (9460) 


1714 (5623) 


2682 (8800) 


- - 


- - 



♦All ATLIT performance predictions were computed assuming that power available = 400 BHP. 

**These single-engine characteristics were combined using a C = 1.05 C to account for engine-out trim drag and using half of the total assumed 
power available. 



An examination of propeller-driven aircraft shows that for 
the higher speeds in an airplane performance envelope, the induced 
drag averages about 10 percent of the total airplane drag. In 
addition, zero-lift drag coefficients for airplanes of a given class 
of aerodynamic sophistication were assumed to be of the same order. 
It was also assumed that propulsion efficiency for a given class of 
airplanes was approximately the same. With these assumptions, maximum 
speed of the airplane is then related to the ratio of wing loaiding to 
power loading 

V ~ y^i . (5.10) 

To estimate climb performance, the basic performance equation 
can be written 



fi = 33,000 (n^ - 



y^/S 



19- (0^3/2/ ) ^ 



) . 



"(5.11) 



Equation (5.11) simply states that rate of climb (li) is proportional to 
the difference between power available and power required. Approximately, 
then, maximum rate of climb occurs at the point on the power curve where. 

3/2 
(C, /r ) is maximum, that is, at the point for minimum power required 

for a propeller-driven aircraft. In practice, best rate of climb does not 



occur at the point for 



fr 3/2 



7, 



J 



; however, for rapid performance 



max 



123 



estimations or for comparative purposes, the assumption is a fair one. 
Actual maximum rate of climb must be computed taking account of the 
changes in propulsive efficiency with forward velocity. The maximum 

rate of climb will then be found where maximum excess power occurs. 

3/2 
For a parabolic drag polar, the maximum value of C, ' /Cr, and 

corresponding lift coefficient can be given as 

0.75 
(r 3/2,. . _ 1.345 (Ae) 

^ ' 
and 

C, = /3 Cp, TT Ae ■ . (5.13) 

'-C "^0 

The airspeed for best rate of climb can be estimated using the lift 
coefficient from equation (5.13). 

For estimating range, the values for the maximum lift-to-drag ratio 
and the corresponding lift coefficient can be given as 



/jrAe 

/c 



fL/'^U^l/^/C,^ (5.14) 



and 



C, =^ AAe Cr, (5.15) 

^M ^0 



The airspeed for best L/D can be estimated using the lift coefficient from 
equation (5.15). 

Single- and multi-engine service ceilings were estimated by solving 
for the density ratio (a) in equation (5.11). With h defined 
(h = 15.2 m/min (50 fpm) for single engine and fi = 30.5 m/min (100 fpm) 



124 



for multi-engine^, the altitude for service ceilings was then computed 
from the resulting value of a. 

Since one of the primary values in this method is for making 
performance comparisons, estimates were computed using both the standard 
Seneca geometry and the ATLIT geometry. These comparisons appear in 
columns (2) and (3) of table 5.4. 

It is instructional to evaluate the ATLIT wing modifications using 
the above equations while varying e and C^ . 



Table 5.5 presents the results of these computations for five cases as 
follows: 

1. Using flight-measured values of e and Cp. for the standard 



Seneca: e = 0.78, C^ = 0.026. 



2. Using the values of e and Cp. predicted for the ATLIT 



design: e = 0.8, C-, = 0.035. 



3. Using values for e and C-, which have been determined 

■^0 

based on preliminary flight- test results: e = 0.67, Cp. = 0.044. 



These values represent the airplane in a configuration with 
massive wing- body interference-induced separation (which 
reduces e), and with added drag due to excessive wing trailing 
edge thickness, instrumentation noseboom, protruding flap and 
spoiler brackets, square wheel well openings, and ten 



125 



126 
inspection covers located along the span protruding into 
the airstream (all of which increase Cp ). 



4. Using values of e and C^ which assume lift-dependent 



drag is cleaned up (i.e., that e is increased) and zero-lift 
drag is unchanged from the value based on flight- test results: 
e = 0.80, Cpi = 0.044. This case demonstrates the effect e 



has on (Cl^^^/Cd) and {L/D)^^. 
max 

5. Using values of e and Cr, which assume zero-lift drag is 



reduced and e is unchanged from the value based on flight- test 
results: e = 0.67, C^ = 0.035. This case demonstrates the 



effect Cp has on {C.^^^/C^) and H/O) • 

max 



126 



TABLE 5.5- COMPARISONS OF PREDICTED PERFORMANCE PARAMETERS 

FOR THE STANDARD SENECA AND ATLIT (W = 1.87 kN. a = 1) 



to 

-3 





Maximum Climb Performance 


Best Cruising Performance 


Case 


by eqn. (5.12) 


Minimum 

Thrust Power Required* 

i.e. @{Cl^^% ) 

^ max 


by eqn. (5, 14) 

(Cl for (L/D)„, -) 


Power Required*** 


1. Piper Seneca 
A = 7.25 

S = 19.4 m^ 
e = 0.78 
Cp = 0.026 




12.3 


70 kw (81 hp) 


13.1 
(0.68) 


69 kw (92 hp) 


2. ATLIT-best case 
A = 10.32 

S = 14.4 m^ 
e = 0.80 
Cq = 0.035 




15.2 


57 kw (76 hp) 


13.6 
(0.95) 


65 kw (87 hp) 


3. ATLIT-worse case 
A = 10.32 

S = 14.4 \/ 
e = 0.67 
Cq = 0.044 




12.,5 


69 kw (92 hp) 


11.1 
(0.98) 


78 kw (105 hp) 



TABLE 5.5- Concluded. 



IN9 
00 





Maximum Climb Performance 


Best Cruising Performance 


Case 


by eqn. (5.12) 


Minimun 
Thrust, Power Required* 

i.e @(C,,^%J 

max 


by eqn. (514) 

(^L ''' «^/^^max**> 


Thrust 

Power Required*^ 


4. ATLIT- 
degraded C^ 



A = 10.32 

S - 14.4 m^ 
e = 0.80 
Cj, = 0.044 




14.3 


60 kw (81 hp) 


12.1 
(1.07) 


68 kw (92 hp) 


5. ATLIT- 
degraded e 
A = 10.32 

S = 14.4 m^ 
e = 0.67 
C|j = 0.035 




13.3 


65 kw (87 hp) 


12.5 
(0.87) 


74 kw (99 hp) 



*From eon ^5 11 V Power required (hp) 
r-rom eqn. ^b.iij. weight (lb) 

for any (C^^/^/Cp) . 

**^(L/D)^^^.C = A Ae C 

M 

***At (L/D) , (C, 3/2/r . _ 
^ ' 'max ^ L f^n) - 



/ws 



19 (Cl^/^/Cjj) ^ 



ro 

00 



-M 



Cpj ^. 1 . C|2 
irAe M 



. Then, power required at (L/D)^^^ is computed by the 

max 

equation above*. 



An analysis of the results in table 5.5 demonstrates two major points 

concerning the relative effects of e and C^ on (C. ' /Cp) 

max 

(the most important parameter for climb performance) and (L/D) 

(the most important parameter for cruise efficiency). 

1. Improving e has stronger effect on (C. ' /C^) than it 

max 

does on (L/D)„,„. A comparison of case (2) with (5) and case (2) 

max 

with (4) shows that improving e from 0.67 to a near-optimum 

3/2 
value of 0.80 reduces (C, /Crj) more than does reducing 

max 
Cj, from 0.044 to 0.035. 



2. Improving C^ has a stronger effect on (L/D)„^^ than on 

Un max 

(C. ' /Cpj). A comparison of case (2) with (4) and case (2) 
with (5) shows that reducing C^ from 0.044 to 0.35 increases 



(L/D) more than does increasing e from 0.67 to 0.80. 

An important observation can be made using the data in table 5.5 
concerning the effect of wing loading on climb performance. Comparing the 
Piper Seneca (case 1) predictions with the ATLIT best-case (case (2)) 

predictions, it can be seen that the value of (C. ' /Cr.) is increased 

max 

by about 27 percent for ATLIT. However, as illustrated by comparing the 

3/2 
values for climb power required, the large increase in (C, /Cr,)„ 

L u max 

for ATLIT does not translate into a 



129 



proportionately large decrease in power required to climb. Thus, the 
predicted improvement in climb performance Cgiven in table 5.4) is small 
(about 12m/min (40 ft/min)). It can be shown that had the aspect ratio 
been allowed to increase while wing loading remained constant, the 
improvement in climb performance over case (1) would have been much larger. 
Cases (3), (4), and (5) demonstrate the potential penalties paid in 
increased power required to climb due to lack of attention to construction 
details, allowing e and C^ to be degraded; that is, not one of these 



cases for ATLIT predicts improved climb performance. 

A final observation which can be made using the data on table 5.5 
concerns the effect of Cj, on the predicted best cruise performance. A 



comparison of the Piper Seneca (case (1)) predictions with the ATLIT-best 
case (case (2)) predictions shows that C-Zp) ,„ is Increased and power 

required at (L/D)„^„ is decreased for ATLIT. Cases (3), (4), and (5) 
max 

demonstrate the penalties paid in increased best-cruise power required 

due to degraded e and C^ ; that is, none of these cases predict 



cruise performance improvements for ATLIT. 

5.4.1.2 Method B: Lift, Drag, and Performance Predictions* 

These predictions, of lift, drag, and pitching moment to be 
encountered during cruise flight, were developed using the computer program 



*The materials for this section were prepared by Dr. Frederick 0. Smetana 
of North Carolina State University. 



130 



described in reference 34 and the vehicle's geometry as obtained from 
Piper shop drawings. Performance predictions were also made using, in 
these instances, the programs described in reference 30. 

LIFT AND DRAG PREDICTIONS 

Wing 

The ATLIT airplane employs a straight, tapered wing with a GA(W)-1 
airfoil section 17-percent thick. The computational technique distributes 
65 regions of constant vorticity on the surface of the airfoil, calculates 
from this an inviscid flow field and pressure distribution, then determines 
the boundary layer growth corresponding to this pressure distribution, and 
recomputes the inviscid flow field of a pseudo airfoil whose ordinates are 
now the physical airfoil ordinates plus the local values of 6* with a 
modification so as to locate the trailing edge stagnation point downstream 
in the wake. This process goes through four iterations so that the 
computed pressure distribution obtained after the last potential (inviscid) 
solution is essentially the same as that used to generate the boundary 
layer solution which formed the basis for that potential solution. The 
program gives section lift, drag, and moment. The drag includes both 
skin friction drag and form drag. However, because of the flow model used, 
extensive regions of flow separation cannot be treated. For this reason, 
the data are unreliable above C, = 0.8. 

The outputs (lift, drag, and moment vs. a for a given Reynolds 
number) from the airfoil program are fed into a curve fitting routine 
which provides polynomial representations of the results for use by the 
wing program. This program uses lifting line theory to modify the local 



131 



angle of attack which the airfoil data "sees" according to spanwise changes 
in twist, camber, thickness, and chord length. Spanwise variations in 
Reynolds number are handled by providing as input tip and root data at 
the correct Reynolds number with the program interpolating to obtain the 
data for other spanwise stations. Inviscid wing-fuselage interference is 
treated by transforming the fuselage mathematically into a vertical slit 
and distributing its effects along the span. The output of the program is 
the three-dimensional lift, drag, and pitching moment of the wing. Note 
that the drag includes both profile and induced drags. 

The same procedure is employed to find the contributions of the tail 
surfaces to the overall aircraft lift, drag, and moment. The vertical 
tail was considered to be half of a symmetric surface unaffected by the 
presence of the horizontal tail. The horizontal tail was assumed to be 
unaffected by the presence of the vertical tail, propeller slipstream, 
or the downwash of the wing. 

Fuselage and Nacelles 

The program to compute the forces and moments on isolated, quasi- 
streamlined bodies having a plane of symmetry represents the half surface 
by 560 flat panels of more or less equal area. On each panel is 
distributed a uniform source whose strength is such that the flow due to 
all sources is everywhere parallel to the surface. Then, a streamline 
which goes through the centroid of a particular panel is traced upstream 
to its inception point. Along this streamline is calculated the boundary- 
layer displacement thickness and skin friction by a momentum integral 
method. This is done for all 560 panels. At the downstream end of the 
body, the wake is arbitrarily assumed to begin at the upstream end of the 



132 



last two sets of panels. The angle of the wake leaving the body is 
determined by the history of the boundary layer displacement up to that 
point. This wake is then paneled to a stagnation point downstream in 
the physical wake and the inviscid pressure distribution on the body plus 
wake body recomputed. The calculated skin friction is integrated over 
the body to find the skin friction drag and the recomputed pressure 
distribution is integrated in the normal and axial directions to find the 
lift and form drag. The same data are also used in computing the 
pitching moment. 

Because the boundary layer routine used is two-dimensional, it is 
not valid when the flow is expanding or contracting rapidly, i.e., near 
the nose or tail of a body, or when there is a significant cross flow, 
i.e., at angle of attack. For this reason, the aircraft drag computation 
is reasonable only in the cruise configuration. In the context of an 
overall drag computation, this is not unduly limiting because the wing- 
drag calculation fails for high angles of attack as well. Several attempts 
were made to extend the angle-of-attack range of the computation, at least 
for ax i symmetric bodies, by using an axi symmetric finite difference boundary 
layer routine in the plane of symmetry in order to locate the lee- side 
separation point and then applying the Allen-Perkins (reference 35) 
technique to determine the normal force. However, the computed separation 
point was not regularly located sufficiently close to physical separation 
point (as found experimentally) to make this approach viable. 

Of course, modeling fuselages and nacelles for the purposes of drag 
computation as isolated bodies ignores interference effects. While it is 



133 



conceivable that the inviscid aspects of interference could be treated 
adequately (and, in fact, have been in many cases), it will require a 
general three-dimensional boundary layer solution to treat the viscous 
aspects adequately. Since such solution techniques will be some time in 
coming, it continues to be necessary to treat these effects empirically. 
Because other approximations in the model can be expected to yield 
uncertainties of the same order of magnitude, no attempt was made to 
account for these effects. 

Protuberances 

No accounting for the drag due to protuberances was made in the drag 
buildup. In general, it is difficult to predict the geometry and location 
of miscellaneous protuberances during preliminary design. The drag due 
to these protuberances can be significant. 

Trim Drag 

The trim drag estimates shown in table 5.6 account only for the drag 
of the horizontal tail which is flying at the lift coefficient required 
to trim the wing-body pitching moment. As pointed out by Mr. R. T. Taylor 
(NASA-LaRC), no account has been made of the additional trim drag due to 
the fact that, with the additional trim lift of the horizontal tail, the 
wing must fly at a new (increased) angle of attack. Thus, total airplane 
longitudinal trim drag includes increased wing drag as well as the 
horizontal tail drag. 



134 



TABLE 5.6 

ATLIT DRAG BUILDUP 
C.G. Q 26.5% MAC 



GO 



%i ng 
(deg) 



-4 

-2 



2 

4 

6 

8 

10 

12 

14 



"Wing 



TRIM 



^D^ \/\ 



TOTAL 



-.134569 


-.003474 


.063437 


.001638 


.259752 


.006707 


.454464 


.011736 


.648280 


.016741 


.841217 


.021724 


1.032682 


.02666 


1.221700 


.031549 


1.405880 


.036306 


1.582697 


.040872 



.001879 

.001852 

.0019772 

.00200 

.00210 

.002184 

.002265 

.002346 

.002424 

.002618 



007316 


.038473 


006045 


.037175 


008349 


.039604 


014132 


.04541 


023051 


.054429 


034891 


.066353 


049506 


.081049 


066789 


.098413 


086509 


.118211 


108327 


.140223 



-.138043 
.065075 
.266459 
.4662 
.66502 
.86294 

1.05934 

1.24324 

1.442 

1.6235 



= c., + c. S.,. + c 



TOTAL ^w ^t ^/^w h'^\ 



fus 



'NACELLE 



= C 



jj + Cp S. ,3 + .0018487 + .01299 + 2 (.00722) 
w t ' w 



= ^n + ^n St/c + .0292787 
^w '^t ^'\ 



= r + c s 



Calculated and Estimated Lift-Drag Polar 

As shown in table 5.6, summing the results of the previous calculations 
yields a drag polar represented by the equation 

Cp = 0.0358 + 0.04056 C^^'^^ (5.16) 

This polar, as indicated previously, does not include the effects of 
flow separations at the higher lift coefficients. In an effort to 
develop a more accurate polar upon which to base performance estimates, 
full-scale wind-tunnel test data on a similar aircraft (reference 36.) 
were examined and fitted by the equation 

Cp = 0.035 + 0.051 C^^ + 0.00138 C^^^*^^ (5.17) 

Plots of these equations are shown in figure 5.12. Note that the two 
curves differ little for C, < 0.8. Above C, = 0.8 it Is to be expected 

that equation (5.17) will more nearly represent the behavior of the ATLIT 
than equation (5.16). Despite the fact that equation (5.16) describes the 
drag of an unpowered airplane and that drag under some conditions of 
powered flight may exceed the drag in unpowered flight, equations (5.16) 
and (5.17) were treated as the probable boundaries for the actual ATLIT 
drag polar. Because of the relatively smaller ATLIT wing area (compared 
with the aircraft tested in reference 36) it is not expected that the ATLIT 
drag will rise as rapidly with increasing C, as it does for the aircraft 
of reference 36. Thus, even if the ATLIT drag in powered flight is 
somewhat greater than in unpowered flight, the drag should be below the 
boundary given by equation (5.17). 



136 



00 

-3 




Figure 5.12- Predicted ATLIT drag polar (method B) and wind-tunnel measured light-twin drag polar, 
(Reference 36). 



PERFORMANCE PREDICTIONS 
The drag polars given by equations (5.16) and (5.17) were submitted 
to the point-performance program described in reference 30 along with the 
thrust horsepower data given in figure 5.13. The latter were derived from 
engine test-cell data and propeller-performance charts. They do not include 
any installation-dependent effects. The data given in columns (4) and (5) 
in table 5.4 represent the output of this program. It will be noted that, 
compared with the original Seneca, only small improvements in rate-of-climb 
and cruise speed are expected. This can be explained by the fact that 
although the airfoil itself offers about a 10-percent improvement in 
L/D at C| = 0.8 (the nominal C. for the climb) the wing is responsible 

for only about 40 percent of the total drag. Overall aircraft drag is, 
as a result, only about 4 percent lower. 



138 



00 
CD 



Power, kw 



I t I I I I I I I I I- 

10 20 30 40 50 60 70 80 90 100 110 120 130 IW 150 160 170 



Hp 



Airspeed V, knots 



Figure 5. 13- Assumed thrust power available for performance prediction by method B. 



5.4.1.3 Method C: Lift, Drag, and Performance Predictions* 

The airplane lift, drag, and performance predictions presented 
here are based on the method of reference 29. The drag characteristics 
of the fuselage, nacelles, and empennage were borrowed from the 
predictions of Method B. The prediction of airplane lift and drag 
essentially consists of the following procedure. With a known 
geometric angle of attack, lift data at the a, RN, and M are obtained 
from tables of 2-D, section characteristics. The initial spanwise 
load distribution is then calculated, followed by a determination of 
the (spanwise) induced angles of attack. With these induced angles of 
attack, a new spanwise load distribution is computed. With a 
satisfactory span loading determined, the 3-D aerodynamic coefficients 
are computed (C^, Cq. , Cq ). 

The section characteristics for the NACA 652.-415 and the 
GA(W)-1 airfoils were used to predict wing lift and drag (with nonlinear 
effects) for the standard Seneca and ATLIT. The airplane lift and 
drag predictions are given in figures 5.14 and 5.15. 

The predictions of airplane performance were made assuming 
available shaft power = 298 kw (400 BHP). Propulsive efficiency was 
estimated in Method B. The calculated curves for thrust power required 
and available appear in figure 5,16. 



*The contributions of Mr. Robert T. Taylor (NASA-LRC) in preparing 
materials for this chapter are gratefully acknowledged. 



140 



-^^ 



2.0 



O' ATLIT - power off, stall approaches 

O ATLIT - measured lift (power for level flight) 




Predictions by Method C 
(engine nacelles not accounted for) 

Seneca - computed (power off) 

ATLIT - computed (power off) 

ATLIT - computed (power on) 
Thrust Coefficients 
T 



^T ^ qs 



^ 0.1 



ot, deg 
-0.5 X 
Figure 5.14- Comparison of measured and predicted lift characteristics for ATLIT and the Seneca. 



4i» 



Predictions by Method C 
1^5-,- — — Seneca 

ATLIT 

1.2-- 

.8" 




Figure 5.15- ATLIT and Seneca predicted airplane drag-polar comparison 
(ideal skin friction, 2-D profile drag used with standard roughness). 



Predictions by Method C 



CO 



300 r 400 r 



ATLIT 



s- 

5 
o 

CL 



to 

x: 
1— 



240- 



kw 



180 



120 



60^ 80 




Airspeed V, knots 



Figure 5.16- Comparison of predicted power required for ATLIT and the standard Seneca. 



With the assumptions above, the Seneca and ATLIT performance 
predictions were computed and are shown in columns (6) and (7) of 
Table 5.4, Without detailed knowledge of the propulsive efficiencies 
involved, the primary value of computations such as these is for 
making comparisons. A comparison of column (6) with column (7) shows 
predicted increases of 9 knots in top speed and 18 m/min (60 fpm) 
in best single-engine rate of climb. The expected ATLIT performance 
would then be determined by adding these increments to the basic 
Seneca performance figures. 



144 



5.4.2 Methods , Dat a Reduction, and Flight-Test Results* 
Speed-Power Relationship 

The variation of shaft horsepower with airspeed was obtained by 
nominally flying the aircraft at constant altitude and airspeed, 
and recording aircraft data with the onboard measurement system. 
Data were averaged over a portion of the record and represent from 200 
to 400 data points. Measured engine RPM and manifold pressure and 
free-air temperature were used with the power-altitude charts of 
reference 39. These charts are for horsepower with the engine 
leaned for maximum power. Since the flight tests were made at a 
rich engine mixture setting, a reduction to rich engine power was 
made using data from reference 40. The resulting value is the brake 
horsepower of the engine. However, since the horsepower delivered to 
the propeller is required to establish the ATLIT performance, the 
power required to drive the engine accessories (alternator, vacuum 
pump, tachometer, propeller governor and fuel pump) as given in 
reference 39 was subtracted from the engine brake horsepower. 

Figure 5.17 presents the variation of horsepower and velocity 
corrected to standard weight and sea level conditions for the ATLIT 
in various test configurations. The basic data points are early 
measurements with the flaps sealed to prevent airflow through the 
spoilers at a flap setting of . Subsequently, a region of flow 
separation on the fuselage and the adjoining wing trailing edge was 
found to exist. Data obtained during the final phase of flight 
tests which resolved the flow separation problem are also plotted on 
figure 5.>7. The aircraft had two fuselage and four wing vortex 



*The contributions of Mr. Joseph H. Judd (NASA-LaRC) in preparing materials 
for this chapter are gratefully acknowl edged - 



145 



250 



Power, kw 



150 



100 — 



50 — 



O 6^ = , gear up 

O" 6j, = , gear up, right strake, vortex generators, tufts 

□ 6^ = 10°, gear up 

n" 5^ = 10 , gear up, right strake, vortex generators, tufts 

A 6, = 30 , gear down 

A fi = 30 , gear down, right strake, vortex generators, tufts 

)^ 6^ = 30 , gear up, right strake, vortex generators, tufts 

















































































- 


























r 


V 




























^ 








, / 


1 


























J 


i^ 






Q^ 




























f 








/ 


!) 










- 
















^- 


i 






o< 


/o 

n 




























/ 






^ 


y 
























-^y. 


\^ 


A 


i 


/Irfi 




V 


p — 






















L 


^ 


/c 


^ 




o 




















- 










lif 


lt3S- 




r 






































































































































- 


































_ 





































400 



300 



250 



hp 



200 



100 



20 40 60 80 100 120 140 160 180 

Airspeed V, knots 



Figure 5.17- Variation of shaft power with airspeed (6^ = 0°, 10°, 30*^) 



146 



generators, a leading edge extension (strake) at the right wing/ 
fuselage intersection (see figure 5.1) and tufts on the fuselage, wing, 
and nacelles. At higher airspeeds (flaps and gear up), an 
increase in power required due to these devices is evident. A 
large portion of this drag increase is caused by the tufts. The 
data on figure 5.17 shows scatter which is attributable to horsepower 
variations. The procedure for obtaining horsepower from charts 
using manifold pressure and RPM is estimated to give values within 
±2 horsepower which falls, within the scatter band. Since most of 
the flights were made between 1000 and 1600 hours, the more likely 
cause of the variations is a larger scale atmospheric motion, 
i.e., data runs were made in slightly rising or sinking air. 
Although altitudes were sought for data runs that did not have 
turbulence (as determined by pilot observation of airplane response 
to external disturbances), large scale atmospheric disturbances 
(long waves) would not be easily detected. 
Aerodyna mic Coeff icients 
The lift coefficients obtained during these runs were computed 
using average values of measured atmospheric conditions and flight-path 
angles and a faired value for aircraft weight. The angle of attack 
and static pressure were corrected for position errors as described 
in chapter 5.1. The variation of power-on lift coefficient with angle 
of attack is shown in figure 5. IS, The increase in lift coefficient 
caused by improving wing-fuselage juncture flow is apparent on these 
figures. Although stall is not shown on these figures, the stall lift 
coefficient is very little higher than the highest C, shown. 



147 



O i5f = 0°, gear up 

0' <Sf = o'', gear up, right strake, vortex generators, tuftS 

□ 6^ = 10°, gear up 

ET 5f = 10°, gear up, right strake, vortex generators, tufts 

A tS^ = 30°, gear down 

A (5j: = 30 , gear down, right strake, vortex generators, tufts 

A 6jr = 30 , gear up, right strake, vortex generators, tufts 




a, deg 



Figure 5.18- ATLIT lift characteristics 
(power for level flight; 6^ = 0°, 10°. 30°). 



148 



The power-on drag of the ATLIT in steady level flight is 
primarily determined by the force developed by the propeller in the 
direction of flight. This force is proportional to the propeller 
shaft horsepower, propeller efficiency, the propeller installation 
efficiency, and a function of the rotation of the propeller force 
vector. An estimate of the propeller efficiency was obtained 
from the Hartzell Corporation, and a computation using the Borst 
method (ref. 41) was found to agree with this estimate. The flow 
field of the nacelle affects the efficiency of the propeller. If 
the propeller is operating in a reduced velocity field -- say in 
front of a bluff body — the apparent propeller efficiency is greater 
than that of the isolated propeller; conversely, if the spinner- 
nacelle geometrically acts to put the propeller in an increased 
velocity region, the apparent efficiency is less than that of the 
propeller. Since the horizontally-opposed engine nacelle combines 
both factors, an installation efficiency of 96 percent was estimated 
using data presented in. reference 42. Experimental data on propulsive 
efficiency of a. wing-nacelle-propeller installation as a function of 
angle of attack and propeller location is presented in references 
43 and 44. An empirical relationship (1 - sin a.^^ ) was found to 
provide a good correlation for the variation of propulsive efficiency 
with angle of attack. Physically, this involves rotating the thrust 
vector so that the propeller slipstream leaves parallel to the mean 
chord line at the trailing edge and subtracting the drag component 
of the thrust vector from the propeller thrust. The expression for 
the drag coefficient then becomes 



149 



(SP)nnni-h - sin^ (a + a.„) x Const 

Cd = ^-^ 5LJ (5.18) 

qSV 

where SP is shaft power. It is estimated that uncertainties in 
these estimates influence the drag coefficients to ± 5 percent over 
the range of test conditions. 

The variation of lift coefficient with drag coefficient is 
shown in figure 5.19. The scatter in the data is primarily due to 
the scatter in measured brake engine power. Note that the calculated 
drag coefficient includes the trim drag and some portion of the 

drag due to power effects, 

2 
The variation of C, with Crj is shown on figure 5.20. 

The curves are apparently nonlinear. The nonlinearity is attributed 

to the effect of power. The variation of C. with C^ due to 

aerodynamic effects is an exponential function of C, whereas the 

variation of the lift due to power will be a trigonometric function. 

The zero-lift drag coefficient of 0.045 for flaps at was obtained 

by extrapolating the data of figure 5.20 to C, = 0. This compares to 

the estimated. value of 0.0358 from table 5.6 at a center of gravity 

location at 26.5 percent mean aerodynamic chord, whereas the measured 

values are for an aircraft center of gravity between 15 and 13 percent 

mean aerodynamic chord. Further, the effect of protuberances was 

neglected in the estimate. A rough check was made to find the order of 

magnitude of the protuberance increment using data from reference 45. 

The estimated equivalent flat plate area of the ATLIT was 0.516 m 

(5.549 square feet), and the equivalent flat plate area of 22 obvious 



150 



O 6^ = 0°, gear up 

Cf d = 0°» Qfi^r up, right strake, vortex generators, tufts 

n 5^ = 10°, gear up 

Uf df = 10°, gear up, right strake, vortex generators, tufts 

A (5-: = 30°, gear down 

^ ^f = 30°, gear down, right strake, vortex generators, tufts 

A 5- = 30°, gear up, right strake, vortex generators, tufts 



4.0 



3.0 



2.0 



1.0 







>"^ 




/ 


- 


yi 


A 


/ 
/ 


/ 

n 
,0 


- 


- — 






?■ 












/ 


■>. 








y 

( 

A 


^/ 




/ 

/ 










/ 












f 


A 










/ 


— 


















Cf 






A 












CX 












> 


?^ 


r 










































— 


— 














































/ 


..__ 






u 








































p- 


— 



























































































0.1 



0.2 



0.3 



Figure 5.19- Power-on, trimmed drag polars for ATLIT (6^ = 0°, 10°, 30°) 



151 



O 6^ - » gear up 

C 6f = 0°, gear up, right strake, vortex generators, tufts 

n 6f = 10°, gear up 

CfSf = 10°, gear up, right strake, vortex generators, tufts 

A 6f = 30°, gear down 

A 6f = 30°, gear down, right strake, vortex generators, tufts 

A 6^ = 30°, gear up, right strake, vortex generators, tufts 



15.0 



10.0 



5.0 



■qj 



T' 




.A-- 



-^ 



-/ 

/ 



(il 



^ 



r 



,Jl 



-h 



/- 



7 






t 



/ 



/ 



^ 



y 



^ 



/^ 



0.1 



0.3 



Figure 5.20- Linearized, power-on drag polar for ATLIT, (6^ = , 10 , 30 ) 



152 



7 2 

items was 0.055 m (0.592 square feet) for a total of 0,570 m 

(6.14 square feet). This compares with the value from measured data 

of 0.648 m (6.975 square feet). Since power effects were also 

neglected in the estimate and would bring estimated and measured 

data closer together, it may be concluded that the method for 

estimating baseline configuration drag by Method B is acceptable. 

However, performance measurements based on these values will be 

optimistic. 

The variation of C, ' /C^ with C^ is shown in figure 5.21. 
The measured value of 11.75 for the flaps-up condition compares with 
estimated values of 12.3 for the Piper Seneca and 12.5 for the worst 
case ATLIT from table 5.5. Removal of the tufts from the ATLIT is 
expected to raise this value. 
Ra te of Climb 

Measured climb data are presented in figure 5.22 for single- and 
multi-engine flight. The multi-engine data are average values at 
610 m altitude (2000 ft.) and at an average aircraft weight of 
1860 kg (4100 lb.). The aircraft had a single strake on the right 
wing, vortex generators on the wing and fuselage, and tufts on the 
fuselage, wings and nacelles. Cowl flaps were closed for the multi- 
engine climb. The figure shows that changes in rate of climb are 
quite small for variations of airspeed from that for best rate of 
climb (approximately 91 knots). Pilot A noted in chapter 5.5.1 
that this climb stability is a desirable airplane characteristic. 
The best rate of climb for the Seneca I is 390 m/min (1280 ft/mi n) 
at the test conditions noted above, and is about the same as that 
measured for the ATLIT. 



153 



O 6^ = , gear up 

C 6f = 0°, gear up, right strake, vortex generators, tufts 

D 6^ = 10°, gear up 

CTSf = 10°, gear up, right strake, vortex generators, tufts 

A 6f = 30°, gear down 

A 6f = 30°, gear down, right strake, vortex generators, tufts 

A 6 -, = 30°, gear up, right strake, vortex generators, tufts 



30,0 














— 


- 














-- 






















































































































































— 
































c 

cr 



















































> 


/ 














r 


























nnf 


cf 






















o 


/ 












ito 


w c 

-i 










nr 
















-■ 








c 


m 












1 
1 


rb 
















- 
























- 


















3/2 



Figure 5.21- Variation of (C^ /%) with lift coefficient for ATLIT 
(5f = 0°, 10°, 30°). 



154 



400 



350 - 



300 



250 
R/C, m/min 

200 

150 

100 

50 



1400 



1200 



1000 



800 



ft/mi n 



600 



400 



200 



90 




right strake on 
vortex generators on 



Flagged symbols indicate turning flight 
with spoilers flush 



One engine 



95 



100 105 110 
mph 



115 



120 



as 



90 



95 



Figure 5,22.- Variation of rate-of-cl imb with airspeed; single-engine 

and multi-engine (W = 17.8 N to 18.2 N (4 000 to 4 100 lbs) 
and 6^ = 0°, 5°. 10°). 



155 



Single-engine rate of climb at 610 m altitude (2000 ft) was 
obtained at an average weight of 1814 kg (4000 lb) with the airplane 
configuration as described above. The operating engine cowl flaps 
were open, while the inoperative engine cowl flaps were closed. The 
ATLIT has no roll trim capability, and straight flight can only be 
obtained by using spoilers on the side with the operating engine. 
To find the penalty involved, the spoilers were set approximately 
neutral, and the airplane was allowed to climb in slowly circling 
flight. A significant increase in rate of climb occurred at 100 kts 
airspeed. It is postulated that most of this increase can be retained 
by use of trim ailerons outboard of the flaps. 



156 



5.5 Pilot Descriptions of Stability and Handling Qualities* 

Two separate pilot evaluations of ATLIT are presented In this 

chapter. Both pilots discuss longitudinal and lateral stability and 

control characteristics throughout the airplane flight envelope; 

Care should be exercised by the reader in interpreting, the ATLIT 

pilot ratings for roll -control tasks. The rigging of the spoilers on 

ATLIT (either down or flush, as described in chapter 3.2.2) strongly 

influences the lateral control feel characteristics of this airplane. The 

purpose in rigging the spoilers symmetrically down into the wing was to 

investigate performance penalties due to spoiler float above the wing. 

Without exception, the pilots reported that the re-rigging greatly degraded 

the lateral handling qualities. The performance changes due to re-rigging 

were negligible. The pilot comments which follow apply to the airplane 

with the spoilers rigging symmetrically down into the wing. Therefore, 

the pilot ratings, in some cases, are excessively harsh compared to what 

they would have been had the airplane been rated with the spoilers rigged 

statically flush (allowing some float at a negligible performance penalty). 

5.5.1 Pi 1 p t A Co mments on ATLIT Flying Qualities 

Cruise Stability 

Cooper-Harper Ratings 
(see table 5.7) 

Spiral - 4 

Longitudinal - 3 

During cruise, it is impossible to fly the ATLIT without constantly 

controlling the aircraft. Small upsets from rough air cause the pilot 



*The contributions of NASA-LaRC research pilots Mr. Robert A. Champine 
and Mr. Philip W. Brown in preparing the materials in this chapter are 
gratefully acknowledged. 



157 



to make corrections in roll and pitch. The spiral stability is weak and 
the spoiler friction is very high, about 44 N (10 pounds) wheel force. 
Since there is no lateral trim control surface, the rudder trim tab must 
be used. Thus, rolling moment due to sideslip is used to trim laterally. 
This lateral trimming procedure may be described as follows. The pilot 
must first look at the wheel or spoilers to be sure they are down flush. 
Then the rudder is moved to maintain the wings level and the rudder trim 
tab is used to reduce the rudder forces to zero. The rudder trim tab has 
a great deal of friction, and along with considerable rudder friction, the 
trimming task is difficult at best. 

TABLE 5.7- HANDLING QUALITIES RATING SCALE 



ADCOUACV con SELECTED TASK OR 
REQUIRED OPCRATION* 



AIRCfUFT KHANDS ON THE PILOT HLOT 

CHAMCTBIISTICS IH ttLECTED TASK OR REQUIRED OPCTATIOM* RATMQ 






Excellent 
Highly desirable 



Pilol compensalion not a factor tor 
desired perlormance 



Good 

Negligible deficiencies 



Pilot compensation not a factor lor 
desired performance 



Fair — Some mildly 
unpleasant deficiencies 



Minimal pilot compensation required lor 
de&ired p^^rformance 




Minor but annoying 
deficiencies 



Desired performance requires moderate 
pilol compensation 



luloderately objectionable 
deficiencies 



Adequate performance requires 
considerable pilot compensation 



Very objectionable but 
tolerable deficiencies 



Major deficiencies 



Major deficiencies 



Major deficiencies 



Major deficiencies 



Pilot decisions 



Coopei-Harpar R«I.NASATND-51S3 



Adequate periorrnance requires extensive 
pilol compensation 



Adequate performance not attainable with 
maximum tolerable pilot compensation. 
Controllability not in question 



Considerable pilot compensation is required 
for control 



Intense pilot cornpensalion is required lo 
retain control 




Control will be lost during some portion of 
required operation 



* Oelinllion al requi'Qd operation involves dnignalion ol fligril phaM and/or 
ftubphases with accoTnpanying condHioni. 



158 



The longitudinal stability in cruise is satisfactory except 
for a slight friction problem. Altitude control is pretty good in 
general. The control force is light, the damping is good, and trimming 
is fairly easy. Also, the phugoid oscillation seems to be of small 
amplitude and of a fairly long period. 

Slow F light a nd Stall Characteristics 



Cooper- Harper 


' Ratings 




Flaps 


up 


3 


Flaps 


down 


$ 


Power 


effects - 


2 



Flaps Up: Slow Flight and Stall 

In general, the ATLIT flys quite well at low speeds, flaps up. 
The roll control is poor for small (up to 25%) roll control inputs 
but is satisfactory at higher deflections. There is plenty of pitch 
control for stalling the wing and also for stall recovery. The rolloff 
at stall is not too bad, being about 20 degrees of maximum bank. The 
stall buffet warning is about 5 or 6 mph, which is good.* During 
the stall, the nose falls through at a modest rate and recovery is 
quick after lowering the angle of attack and increasing power. Recovery 
can be effected without losing more than 15 m (50 feet) of altitude. 
The power-off stalling speed is fairly high, about 70 knots (80 mph); 
however, if proper operating procedures are adhered to, this presents 



*This stall buffet warning disappeared with the addition of the devices 
for attacking wing-body interference-induced flow separation (i.e., 
strakes and vortex generators). 



159 



no problem. By this it is meant that anytime one wishes to fly at 
speeds below about 96 knots CllO mph), the flaps should be extended 
between 5 and 10 degrees to increase the stall margin. 

Flaps Down: Slow Flight and Stall 

In general, the flaps -down slow-flight characteristics are 
pretty good. The roll response is good at all speeds, but the roll 
system friction and force gradient near center are yery bad (this is 
the main reason for poor Cooper-Harper rating). At minimum speeds, 
particularly with flaps set at 30 or 37^ degrees, the longitudinal 
stability and damping are very weak. The pitch control is still very 
responsive. This can lead to some overcontrolling in rough air during 
landing. During stall recovery, this overcontrolling can also be a problem. 

In general, stall characteristics with flaps down are good with 
little or no rolloff. The spoilers are very effective throughout the 
stall, and recovery can be made with little loss in altitude. Deep 
stalls, using full back-pitch control, have not been investigated, and 
no comment can be made at this time. Stall warning is in the form of 
airframe shaking; in fact, one can look at the horizontal tail surface 
and see it shaking up and down about ±1.3 cm (+1/2 inch) at the tips. 
There has been a wing-fuselage separation that buffets the tail and 
provides about 5 knot warning before the stall. All stall warning seems 
to be eliminated since the wing-fuselage separation was cleaned up with 
strakes and vortex generators. 



160 



Multi- and Single-Engine T ri m with Power Changes 

Cooper-Rarper Rating 

Multi-enginjS 2 
Single-engine 6 
Power effects in the ATLIT are very good. The addition of power 

causes a normal nose-up trim change. The trim system is fully capable 

of zeroing the forces due to the power- induced trim changes. The control 

force can be controlled with one hand during a go-round. Since the 

propellers counter- rotate there is no torque effect. 

The single-engine performance is yery. marginal, and this is 

the reason for the 6 Cooper rating. At low speeds (below about 

96 knots), there is not enough rudder trim authority to trim out rudder 

forces. At speeds down to 78 knots (90 mph), two feet on one rudder 

pedal are required. The force is very high. The spoilers are effective 

in controlling the single-engine forces,, but if they are raised more 

than about 1 cm (3/8 inch), then the additional drag degrades the 

marginal single-engine climb performance. 

Rolling Performance with Spoilers 

Maximum Rolling Performance 

Cooper-Harper Rating 
Flaps UP - 3 
Flaps down - 2 



Lateral Control Feel 



Cooper-Harper Rating 
Flaps UP - 4 
Flaps down - 6 



161 



The maximum (full wheel: deflection) rolling performance with 
the spoiler control is excellent at all speeds and flap deflections. The 
rolling velocity is very low when small wheel deflections (up to 255^ 
of total J are used. When 50% of total wheel deflection is used. the 
rolling velocity is more than adequate for most flight conditions. This 
nonlinearity is mildly unpleasant and is something with which the 
pilot has to cope. When using small spoiler deflections, there is a 
small adverse yaw before actual turning flight is started. However, 
when using more than about 25% deflection proverse yaw results and 
turning flight starts immediately. These characteristics are very good. 
These comments apply to all flap conditions. 

The lateral control feel forces are poor because of very high 
friction and a negative centering force gradient up to about 25% 
deflection. These negative forces, off from center, are greater when 
the flaps are down. These unacceptable forces need correction, as does 
the friction level. When the flaps are retracted, the wheel centering 
forces are nearly zero but the friction is still high. 

Crosswind Landings (Sideslip Characteristics) 

C ooper-Hanper Ratings 

Sideslip - 2 

Crosswind 
landing's - 5 

The sideslip characteristics are good, as the airplane will fly 

at fairly large sideslip angles. Rudder and spoiler effectiveness are 

good, and no unusual pitching moments have been noticed. 



162 



Crosswind landings are another matter because precise control is 
required. Usually, crosswind conditions involve gusty and changing 
directions of the wind. Under these conditions, control of the sideslip 
angle is difficult and unpleasant. This is due to the nonlinear roll 
response with small spoiler deflections. Also, the poor force gradient 
(negative centering) and high friction add to the problem. There is a 
tendency for overcontrolling due to the low rolling effectiveness at small 
wheel deflections followed by good roll response at approximately 25-percent 
deflection of the wheel. Therefore, during gusty wind conditions, the pilot 
must rapidly move the wheel right and left through ±25 degrees of travel to 
counter the shifting winds. At best, it can be said that adequate control 
is available, but a very high skill level is required to make a good 
crosswind landing during gusty wind conditions. 

Instrument Approaches 

No ILS-type instrument approaches have been made by this pilot and 
comment is only conjecture. Control of the aircraft on the ILS approach 
would be unsatisfactory because of roll-control friction, poor centering 
forces in roll, lack of trim in roll, and nonlinear roll response. These 
items have been discussed above. 

The reader should note that the poor Cooper ratings for lateral feel 
characteristics apply to the ATLIT with the spoilers rigged symmetrically 
down below the wing surface, as described in chapter 3.2.2. This rigging 
not only worsened the roll response for small wheel deflections (i.e., it 
takes more wheel deflection to raise the spoiler to a positive deflection), 



163 



but the wheel forces also became more wheel decentering. The pilot 
agrees that with the spoilers rigged in the flush position, the poor 
Cooper-Harper ratings would improve by about two grades. 



164 



5.5.2 Pilot B Comments on ATLIT Flying Qualities 

The purpose of this report is to give a qualitative assessment of 
the ATLIT's handling qualities and to assign Cooper-Harper pilot ratings 
for a variety of tasks. 

The quantitative measures used were taken from panel instruments or 
instrumentation package "quick-look" records. Though only approximate 
values, their inclusion is justified to better define the characteristics 
discussed. 

Unless otherwise noted, this report refers to the ATLIT's current 
configuration of vortex generators, leading edge strakes, taped 
protuberances, and clay and balsa-filled recesses. The average weight 
and C.G. are 18.2 N (4100 lb) and 15.5% mac, respectively. 

The following configurations and flight conditions were examined: 



Configurations 
or maneuver 


Landing gear 
position 


Flap 
position 


V trim 
IAS, kts 


Power 


Cruise 


Up 


0° 


139 


a = 


Approach 


Up 


30° 


65 


a = 0; idle 


Stall 


Up 
Up 
Up 
Up 
Up 


' 0° 
10° 
20° 
30° 
37° 


80 
58 
54 
54 
54 


a = 0; idle 
a = 0; idle 
a = 0; idle 
a = 0; idle 
a = 0; idle 


Precision 
heading, 
vertical 
S patterns 
^cruise) 


Up 
Down 


0° 


104 


152 m/min (500 fpm) 
climbs and descents 


Landing 


30° 


65 


Idle 



165 



CONTROL SYSTEM CHARACTERISTICS 

The longitudinal control system utilizes some mass imbalance 
but no downspring. An anti-servo stabilator tab is used for force 
tailoring and trimming. The spoiler system utilizes springs to' provide 
a centering tendency. There is no lateral trim system. Rudder force 
tailoring and trimming are provided by an anti -servo tab. 

Friction + breakout forces 





Cruise 


Approach 


Longitudinal 
wheel Fj. 
force X 


21 N (4.75 lb) 


24 N (5.5 lb) 


Lateral p 
wheel W 
force ^ 


53 N (12 lb) 


36 N ( 8 lb) 



Control Centering 

1. Cruise .- The control wheel will quickly return to 30 to 40 percent 
of the longitudinal-control input necessary for a 1.5 g pulse. Then, in 
5 to 8 seconds, the wheel will creep slowly back toward trim another 
10 percent of the input amplitude. 

Lateral centering is also poor. For wheel deflections of less than 
25 to 30 degrees, there is no centering tendency; large deflections will 
return to this 25- to 30-degree position when the wheel is freed. 



166 



2. Approach.- -AFy MSq-rnn appears to be lower here than in 

the cruise case. Centering is correspondingly worse. 

Lateral control centering tendencies are nonexistent. In fact, 
there is a range of motion on either side of the control wheel centered 
position, out to ±20° of wheel rotation, where AF,. /A6 < 0. The 

control wheel will not actually decenter when freed, however, because 
of the high level of friction present in the control system. 

Control Raps 

Longitudinal control wheel raps resulted in one small amplitude 
overshoot of the final control position. There was no separately 
distinguishable aircraft response to this small overshoot. 

Control Surface Trimming 

The longitudinal electric trim is a little slower than desirable. 
Manual rudder trim is quite satisfactory. The lack of a lateral-trim 
system is considered unsatisfactory because of the necessity to deflect 
the spoilers, sideslip the aircraft, or use differential power to effect 
lateral trim. 

LONGITUDINAL CHARACTERISTICS 

Static 

Both the wheel position and wheel force versus speed relationships 
indicate that positive stick-fixed and stick-free longitudinal static 
stability exist. The force-speed gradient is shallower in the approach 



167 



than in the cruise configuration. The high-control system friction 
results in a wide trim-speed band. Typical figures are given below; 



Cruise Approach 

Trim 

speed band, 139 - 135 kts 70 - 54 kts 
IAS 



Power effects were noted at 104 KIAS, gear and flaps retracted; 
to maintain airspeed from a level -flight power setting to a full -throttle 
climb required F^ = 27 - 31 N (6 - 7 lb). 



Dynamic 



Cruise Approach 



'phugoid 0.1 0.35 



Short period behavior in cruise was sufficiently high frequency 
and well damped enough to allow accurate tracking in pitch. The approach- 
configuration short period was not quite so good; attempts to reset 6 
in a step-like fashion resulted in a one-half cycle overshoot of the 
desired value. 

Maneuvering 

Control wheel position and force versus a indicated apparently 
positive stick fixed and stick free maneuvering stabilities. The 
influence of mass balance in the control system on wheel forces is 
unknown. 



168 



Sinusoidal stabilator inputs across a wide frequency range showed 
no tendency to develop pilot-induced oscillations. 



LATERAL-DIRECTIONAL CHARACTERISTICS 



Static 



Dihedral effect is moderately positive in both the cruise and 
approach configurations. Steady heading sideslips showed the stability 
derivatives involved to have conventional signs. In the approach 
configuration, a maximum steady heading sideslip maneuver resulted in 
the following values: 

3 = 16.3° 

6 u 1-30° 
wheel 

Dynami c 

Spiral .- Spoilers and rudder were held fixed while checking this 
mode and the stabilator was used as necessary to maintain airspeed. 
Lateral movement of a weight within the cabin upset and then reset the 
ATLIT's rolling moment equilibrium. 

The spiral mode is neutral in cruise. In the approach configuration, 
time to double amplitude is about 8 seconds. 

Dutch Roll. 





Cruise 


Approach 


*/3 


0.75 


0.90 


^D.R. 


0.16 


0.16 



169 



STALL BEHAVIOR 

Investigation of stall behavior was limited to cases where a, - 1 

and V £ 1 kt/sec. During the stall, a definite g break and nose 
down pitching tendency occurred. Wing dropping tendencies were mild and 
no tendency to roll in a particular direction existed. With the exception 
of the stall with stabilator stall, stalling behavior is docile. 

Warning 

The original configuration (which lacked the leading edge strakes 
and vortex generators) produced a ^ery vigorous pre-stall buffeting 
of the stabilator. In the present configuration, however, warning of 
impending stalls is practically nonexistent. Buffet onset never comes 
more than 1 to 2 knots before the stall and when present, it is barely 
perceptible. Typically, a very light buffet occurs simultaneously with 
the g break itself. In one case, where 6^ = 10 and power was set 
for level flight, some slight lightening of control wheel pull force 
occurred just prior to the stall. 
Control Feel and Effectiveness 

The rudder remains very effective throughout the stall. Later control 
is best achieved through a combination of rudder and spoiler deflection, 
although in the flaps-up case, the spoilers are nearly ineffective.* The 



*It should be noted that the addition of wing-root strakes and vortex 
generators reduced indicated flaps-up stall speeds by about 9 knots from 
the stalls described by pilot A. The spoiler effectiveness during 
these slower stalls is reduced accordingly. 



170 



stabilator effectively controls a except when the stabilator stalls 
during the recovery from a wing stall. This unusual condition is 
described in more detail shortly. 

Control -position force gradients seem to remain approximately the 
same during stall except for the 6^ = 10 case mentioned under the 

"Warning" discussion above. 

Recovery Technique 

For the original configuration, recovery from all stalls could be 
effected by allowing the stabilator to move slightly off the negative 
stop. Holding the control wheel full aft would result in a moderate 
porpoising motion. 

Recovery technique during the one buffet-onset investigation flight 
in the present airplane configuration consisted of an ^expeditious 
increment of forward wheel movement, followed by an aft repositioning to 
a point corresponding to a higher than stall trim speed. This technique 
quickly unstalled the wing and was satisfactory except in the case of 
stall with stabilator stall. 

Power Effects and Stall Speeds 

Two conditions, idle power and power for level flight, were explored, 
Stall warning was lacking in both conditions. Lateral control was 
roughly the same for both conditions. The stalling speeds were 
significantly affected by power. 



171 







^f 

0° 


10° 


20° 


30° 


37° 


Approximate 


Idle power 


70 


60 


57 


54 


53 


Stan 
speeds 

V ' , knots 
c 


Power for 
V= 


64 


56 


53 


49 


47 


Stan with Stabilator 


Stan 













This phenomenon has been observed only with a 6jr = 37 and power 

for level flight. Figure 5.23 compares stalls with and without stabilator 
stalls. Stabilator positioning after the g break is very similar for 
both stalls. The period necessary to reduce a to the pre-g break 
value was also the same. 

Following the normal stall, flap retraction was begun after it was 
apparent that recovery from the maneuver was in progress. Although this 
was only 2 seconds after the g-break, Q had peaked negatively and then 
reduced in magnitude by 35 percent. 

Similarity between the normal and stabilator stall cases ceases 
2 seconds after the g break. A sudden force reversal occurs and 
V , a', a and continue their divergence. At this point, a 
decision was made to reduce the negative pitching moment by raising the 
flaps. A maximum e of -56° was attained before the pitch divergence 
was stopped. 

Stabilator stall and the resulting pitch divergence was encountered 
on one other occasion, this case occurring with the original aircraft 
configuration (no strakes or vortex generators). Power was set for 
level flight and 6^ = 37°. The trim speed, however, was 65 KIAS, a 



172 



Airspeed 
', ki 



^' I knots 150 -r 



Angle of Attack Ofi- 
a, deg 

+10 



Normal Acceleration 
a^. 9 2 



Pitch Rate 

Q, deg/sec 



Pitch Attitude 
8, deg 



StabHator Deflection 



Longitudinal 

Wheel Force 

F„ . N 



+ 100 ^1 



Flap Deflection ^^ ^ 




z 4 6 8 10 12 14 16 18 20 Z2 



Time, t, sec 

(a) Normal stall 
throttles idled 




2 4 e 8 10 12 14 16 18 20 22 



Time, t, sec 

(b) Stall with stabilator stall 
power for level flight 



Figure 5.23- ATLIT stall time histories. 



173 



value well above the stall speed. As the nose-down portion of a 
stabilator doublet was initiated, a divergent pitching motion developed; 
in that instance, recovery was effected by raising the flaps. 
Unfortunately, no records were taken of that maneuver. 

TASKS FOR PILOT RATINGS 
Takeoff and Transition to Climb 

With a takeoff flap setting of 10 degrees, takeoff trim and a 
somewhat reduced power setting (to prevent too fast an acceleration through 
the speed of interest), the nosewheel could be lifted clear of the 
runway at an IAS = 38 kts. The attitude was then reset, full throttle 
was applied and liftoff was made at 74 kts. Care had to be used to avoid 
an overrotation. Directional control was easily maintained. With the 
gear and flaps retracted, climb power, a moderately heavy push force on 
the wheel was necessary to maintain 87 knots. The electric stabilator 
trim was somewhat slow for coping with the large pitching moment changes 
due to flap retraction and extension. A pilot rating of 2.5 was assigned 
to this flight phase. 

Cruise 

Two IFR tasks were evaluated with and without turbulence. The 
first consisted of precisely holding a heading in level flight. The 
next was the vertical S pattern depicted in the following figure. 



174 



}\) = 3°/sec 




Vertical S Pattern (V = const.) 



Without turbulence, precision heading holding was very easy. In 
rough air, however, excitation of the dutch-roll mode made the task 
difficult. Because of the high control system breakout and friction 
forces, it was difficult to reset the controls to trim after correcting 
for turbulence-induced upsets. Thus, the ATLIT would soon roll off in 
a direction corresponding to the control surfaces' out-of-trim positions. 
The net result was a mildly oscillatory rolling and yawing which was not 
eliminated even with considerable pilot effort. With turbulence, the 
cruise precision heading hold task was given a pilot rating of 4.5. 

The vertical S pattern in turbulence incorporated all of the same 
difficulties encountered with the heading hold task. Additional difficulty 



175 



was introduced by the longitudinal control repositioning necessary as a 
consequence of power changes. The high control system breakout and 
friction forces were largely responsible for the assignment of a 5.5 pilot 
rating for task with turbulence. Without turbulence, there was a slight 
improvement to a rating of 5.0. 

Cruise Turns, Coordinated and Two Control 

Coordinated turns and turn reversals were easily accomplished. Rolling 
performance is indistinguishable from an aileron equppped aircraft. There 
was no noticeable nonlinear roll response to control-wheel inputs. 

Rudder only turns and turn reversals were accomplished in a quick 
and relatively precise manner. Both spoiler-only and rudder-only turns 
and turn reversals are satisfactory alternate methods of lateral control. 
This entire flight phase is assigned a pilot rating of 2. 

Formation Flying 

High control system friction and breakout forces make the ATLIT 
very tiring to fly in formation. As a formation lead aircraft, the 
aircraft's turbulence response coupled with the pilot's corrections for 
upsets leads to' a "wallowing" motion. This phase is given a pilot 
rating of 5. 

Approach 

Precision heading holding was again easy except in turbulence. The 
same type of general aircraft behavior was noted here as in cruise; 
however, the ratio of rolling to yawing disturbances was higher here 
than for cruise. Vertical S patterns were more difficult because the 



176 



lower longitudinal force gradients coupled with control system friction 
and breakout forces increased the difficulty of making corrections based 
on control feel rather than displacement. The previous pilot ratings 
for the precision heading hold and vertical S pattern tasks are 
increased numerically by A = 0.5. 

Nonlinearity of roll response to wheel deflection became noticeable 
in the approach configuration. On gusty days, tight lateral control of 
the aircraft was impossible if control wheel deflections were limited to 
<_ ± ^5^ . It was not unusual to contact the lateral wheel stops (at a 
deflection of 90°) when trying to closely control bank angle. The 
lateral decentering moment "assisted" the pilot in an annoying manner. 
It should be stressed that the maximum roll rate the pilot can command 
is satisfactory. 

For the approach task in turbulence, the pilot rating was 6. 

Approach Turns, Two Control 

Spoiler only turns and turn reversals showed generally the same 
characteristics as in the cruise condition. Rudder only turns from 
(f) = were also similar. But large input, rudder-only turn reversals 
from some 4> i' were quite peculiar; bank angle built up in the 
direction of the rudder input but ^ built up in the opposite direction. 
Tp did not swing back in the direction of the input until the aircraft 
rolled through the wings-level position. Roll attitude control, spoiler 
only, was rated a 3. Rudder only control was given a 6. 



177 



Landing 

The original ATLIT configuration was landed with the wing-low 
method in an 18 knot crosswind. The rudder was frequently on the stop 
for this landing, but, since not much control wheel deflection was 
necessary to achieve this sideslip, the additional wheel deflection 
necessary to counter turbulence upsets was available. 

A recent steady heading sideslip test indicated that a steady 
crosswind component of 18 kts could be handled. 

The flare maneuver could be easily overcontrolled because of the 
difficulty of feeling out the inputs thus, corrections had to be made 
by judging their adequacy initially in terms of displacement rather than 
force. Friction and breakout masking of a stabilator trim position was 
very detrimental here. 

Once on the ground, directional control via the rudder was very 
satisfactory, but the spoilers appeared to be very ineffective for 
directional control (with 6 = 30°). A pilot rating of 4 was assigned 
to the landing phase. 



178 



CHAPTER 6 
CONCLUSIONS AND RECOMMENDATIONS 

6.1 Conclusions 

The conclusions for the ATLIT evaluation presented here are based on 
complete flight-test results of the stalling and the rolling characteristics 
and partial results for the cruise and climb performance. 

1. The stalling speeds and the maximum-lift coefficients were in 

good agreement with the design estimates and the wind-tunnel 

predictions. The stalling characteristics are described by 

the pilots as gentle with wery moderate roll off and adequate 

lateral control throughout the initial stall departure. The 

stalling speed with flaps defTected 37 degrees was 51 knots 

(59 mph) and the corresponding C. was 3.03. With 

max, A 

flaps up, the airplane stalled at 68 knots (78 mph) for a 

corresponding C, of 1.7. This flaps-up maximum 
max, A 

lift and the great effectiveness of the Fowler flap in increasing 

maximum lift are apparently unequal ed for general aviation 

airplanes of a similar configuration. 

2. The spoiler roll-control power met design expectations and 
was in good agreement with wind-tunnel results. In the 
current landing configuration (6^ = 30°), the maximum 

helix angles are greater than 0.11. With the flaps deflected, 
the spoiler roll control on ATLIT exhibits the desirable 
behavior of increasing helix angles with decreasing airspeeds. 



179 



This feature gives the pilot increasing bank-attitude 
control as the airplane slows down during the landing flare. 
No adverse yaw wasmeaslured during rolls; in fact, a small amount 
of proverse yaw was noted with large spoiler inputs. 

Although the spoilers provided very powerful roll control, 
this control system did have undesirable control feel 
characteristics, depending on whether the spoilers were rigged 
up, flush, or down. These feel characteristics result from the 
large amount of combined control -system breakout force and 
friction of about 40 N in combination with the reduced C-, 

for small spoiler deflections. 

Much has been written (references 37 and 46) recommending 
against the use of spoilers alone for roll control. Past 
researchers have endorsed the use of a small trim or feeler 
aileron, along with roll control spoilers, to provide better 
feel characteristics, more positive control for small spoiler 
inputs, as well as to function as a lateral trimming device. 
In light of the experience with the ATLIT spoiler roll-control 
system, this recommendation is still a good one. It is a good 
recommendation not because spoiler systems cannot be designed 
to adequately serve as the sole means of airplane roll control, 
but because the use of a trim aileron greatly simplifies the 
design and implementation of roll spoilers. This situation 
may change when design data for spoilers are available to the 



180 



same extent as for ailerons, thus making the design for a 
mechanically-actuated, spoiler roll -control system as 
straightforward as it presently is for an aileron system. 
3. In the configuration tested and reported on here, neither 
cruise nor climb performance of ATLIT met the design 
expectations. Top speed for ATLIT was 168 knots (193 mph) 
and maximum rate of climb was approximately equal to that 
for the standard Seneca. The most reliable predictions for 
ATLIT performance increases, over the standard Seneca indicated 
about 9 knot (10 mph) increase in top speed to V = 178 knots 

(205 mph) and an increase in maximum single-engine rate of 
climb of about 12 m/min (40 ft/mi n) from about 58 m/min 
(190 ft/mi n) to 70 m/min (230 ft/min). These small predicted 
performance improvements were not realized because of the 
poor span efficiency factor (e - 0.65) and the high value for 
zero lift drag (C^ - 0.045). Calculations showed that with 



proper attention to construction details on the airplane, the 
predicted cruise and climb performance could be realized. 

6.2 Recommendations 

Flight testing of ATLIT in its present configuration (strakes and 
vortex generators on both wings) will continue to the completion of 
climb performance testing. After that, the following three major phases 
of testing are planned. The discussion for each phase includes 
recommendations for topics of pertinent research. 



181 



6.2.1 Supercritical PropalTer Evaluation 

Before the supercritical propellers are installed on ATLIT, baseline 
data on noise and performance characteristics will be gathered with the 
standard propellers. Identical tests will then be done with the 
supercritical propellers on ATLIT. 

Both interior and exterior noise measurements will be made. Exterior 
noise characteristics will be documented by the guidelines of Federal 
Aviation Regulations (FAR) Part 36 (Noise Standards: Aircraft Type 
and Airworthiness Certification). 

Propeller performance characteristics will be measured by constant 
altitude, level flight accelerations, and takeoff distance measurements 
(during these tests, landing distances will also be measured). 

6.2.2 Preliminary Plans for ATLIT Full -Scale Wind-Tunnel Te sts 
ATLIT is scheduled to enter the LaRC full-scale (30- by 60- foot) 

wind tunnel in the fall of 1976. Several possible areas of research 
during these tests are listed below. The research items listed include 
items presented by a poll of U.S. general aviation manufacturers. 
The tentative areas of investigation follow: 

1 . Documentation of baseline aerodynamic and performance characte ristics 

2. Drag cleanup . - Several items have been identified as candidates 
for modification in a drag cleanup program as follows: 

(a) reduce wing trailing edge thickness 

(b) fair flap brackets, spoiler hinges, and other miscellaneous 
proturberances 



182 



(c) construct flush inspection covers to replace 16 presently 
protruding inspection covers located spanwise on wing 
lower surface 

(d) remove instrumentation noseboom 

(e) evaluate improved wheel well fairings 

(f) improve the fit and sealing of the cabin door 

(g) optimize devices for attachment of the wing-body 
interference- induced flow separation. 

3. Cooling drag studies . 

4. Studies of propeller/nacelle interference effects on propulsive 
efficiency and drag . 

5. Studies of trim drag in single-engine climb configurations . 

6 . W ing-wake surveys for document at ion of section profile drag and 
comparison with 2-D results . 

7. Boundary- layer profile measurements for comparison with 2-D data . 

8. Spoiler effectiveness and hinge-moment measurements for 
comparison with 2-D and 3-D scaled data . 

9. Static stability derivative measurements . 

10. Measurements of high angle-of-attack characteristics . 

11. Acoustic (propeller, engine, and airframe) studies . 

12. Evaluation of winglets on ATLIT. 

Obviously, not all of these areas for research can be studied during the 
short time ATLIT will be in the tunnel. An order of priorities remains to 
be determined. 



183 



6,2.3 Final Flight Evaluation 

ATLIT will return to flight status after the full-scale wind-tunnel 
tests. Flight data will be gathered to verify the wind-tunnel optimized 
ATLIT configuration (i.e., to measure, in flight, the effects of any 
fairing or fillet devices or winglets which may be tested in the tunnel). 

Development should continue of the method discussed in appendix A 
for extracting aerodynamic drag and power parameters from flight data. The 
major emphasis in the continuing development of this method should be on 
improved flight-data quality. Additional, independent approaches to such 
a method should be encouraged. 



184 



REFERENCES 



1. Crane, Harold L.; McGhee, Robert J.; and Kohlman, David L.: 

Applications of Advanced Aerodynamic Technology to Light Aircraft. 
Paper 730318, SAE National Business Aircraft Meeting, Wichita, 
April 1973. 

2. Roskam, Jan; and Kohlman, David L.: An Assessment of Performance, 

Stability, and Control Improvements for General Aviation Aircraft. 
Paper 700240, SAE National Business Aircraft Meeting, Withita, 
March 1970. 

3. Kohlman, David L.: Flight-Test Results for an Advanced Technology 

Light Airplane Wing. Paper 740368, SAE National Business Aircraft 
Meeting, Wichita, April 1974. 

4. Raisbeck, J. D. : Consideration of Application of Currently 

Available Transport-Category Aerodynamic Technology in the 
Optimization of General Aviation Propeller-Driven Twin Design. 
Paper 720337, SAE National Business Aircraft Meeting, Wichita, 
March 1972. 

5. McGhee, Robert J.; and Beasley, William D.: Low-Speed Aerodynamic 

Characteristics of a 17-Percent Thick Airfoil Section Designed 
for General Aviation Applications. NASA TN D-7428, 1973. 

6. Stevens, W. A.; Goradio, S. H.; and Braden, J. A.: Mathematical 

Model for Two-Dimensional Multi -Component Airfoils in Viscous Flow. 
NASA CR-1843, 1971. 

7. Garrett, Robert B.; Experimental Investigation of High-Lift Devices 

for a Light Aircraft. M.S. Thesis, University of Kansas, 1972. 

8. Sapp, Charles W. : Application of Spoilers to Light Airplanes. 

M.S. Thesis, University of Kansas, 1969. 

9. Agler, Rex D.: Experimental Investigation of the Influence of Wing 

and Spoiler Effectiveness for Light Aircraft. M.S. Thesis, 
University of Kansas, 1970. 



185 



10. Colwell, Robert C: Improvement of the Performance, Stability, 

and Control of a Current Light Aircraft. Masters Thesis, 
University of Kansas, .1970. 

11. Kohlman, David L.; Holmes, Bruce J.; and Crane, Harold L.: 

Preliminary Flight-Test Results of an Advanced Technology Light 
Twin-Engine Airplane. Paper 750525, presented at the SAE 
National Business Aircraft Meeting, Wichita, Kansas, April 1975. 

12. Holmes, Bruce J.; Kohlman, David L.; and Crane, Harold L.: 

Preliminary Flight-Test Results of an Advanced Technology Light 

Twin-Engine Airplane (ATLIT). Paper 760497, presented at the 

SAE National Business Aircraft Meeting, Wichita, Kansas, April 1976. 

13. Roskam, J.; Kohlman, D. L.; and Wentz, W. H.: Spoilers for Roll 

Control of Light Airplanes. Paper 74-851, AIAA Mechanics and 
Control of Flight Conference, Anaheim, California, August 1974. 

14. Wentz, William H. , Jr.; and Seetharam, H. C: Development of a 

Fowler Flap System for a High-Performance General Aviation Airfoil. 
NASA CR-2443, December 1974, 

15. Wentz, William H., Jr.; Seetharam, H. C; and Calhoun, John T.: 

Wind-Tunnel and Flight Development of Spoilers for General Aviation 
Aircraft. Paper 750523, SAE National Business Aircraft Meeting, 
Wichita, Kansas, April 1975. 

16. Wentz, William H. , Jr.: Effectiveness of Spoilers on the GA(W)-1 

Airfoil with a High-Performance Fowler Flap. NASA CR-2583, 1975. 

17. Wentz, William H., Jr.; and Volk, C. G. , Jr.: Reflection-Plane 

Tests of Spoilers on an Advanced Technology Wing with a Large 
Fowler Flap. NASA CR-2696, 1976. 



186 



18. Smetana, Frederick 0.; and Summey, Delbert C: Drag-Analysis 

Methods for Light Aircraft. Paper 750526. SAE National 
Business Aircraft Meeting, Wichita, Kansas, April 1975. 

19. Ellis. David R.; and Tilak, Narayan W. : An In-Flight Simulation 

of Lateral Control Nonlinearities. NASA CR-2625, November 1975, 

20. Paulson, John W., Jr.: Wind-Tunnel Investigation of a Fowler Flap 

and Spoiler for an Advanced General Aviation Wing. NASA TN D-8236. 
1976. 

21. Paulson, John W. , Jr.: Wind-Tunnel Test of a Conventional Flap and 

Aileron and a Fowler Flap and Slot-Lip Aileron (Spoiler) for an 
Advanced Technology General Aviation Wing. Paper 750501, 
SAE National Business Aircraft Meeting, Wichita, Kansas, 1975. 

22. Weick, Fred E.: Aircraft Propeller Design. McGraw Hill, New York, 

1930. 

23. Roskam, J.: Overview of Trim Drag. Proceedings of the NASA/ Industry/ 

University General Aviation Drag Reduction Workshop, Lawrence, 
Kansas, July 1975. 

24. Fischel, J.; and Ivey, M. F.: Collection of Test Data for Lateral 

Control with Full-Span Flaps. NACA TN 1404. 1948. 

25. Weick, F. E.; and Wenzinger, C. J.: Preliminary Investigation of 

Rolling Moments Obtained with Spoilers on Both Slotted and 
Plain Wings. NACA TN 415, 1932. 

26. Gracey, W.: Measurement of Static Pressure on Aircraft. NACA 

TR 1364, 1957. 

27. Thompson, Floyd L.: The Measurement of Airspeed of Airplanes. 

NACA TN- 616. 1937. 

28. Richardson, Norman R.; and Pearson, Alkin 0.: Wind-Tunnel 

Calibrations of a Combined Pi tot-Static Tube, Vane-Type 
Flow-Direction Transmitter, and Stagnation Temperature Element 
at Mach Numbers from 0.60 to 2.87. NASA TN D-122, 1959. 



187 



29* Sivells, James C; and Westn'ck, Gertrude C: Method for Calculating 
Lift Distributions for Unswept Wings with Flaps or Ailerons by 
Use of Nonlinear Section Lift Data. NACA TR 1090, 1952. 

30. Smetana, Frederick 0.; Summey, Delbert C; and Johnson, W. D.: 
Point and Path Performance of Light Aircraft, A Review and 
Analysis. NASA CR 2272, 1973. 

31., Hoak, D. E.; Ellison, D. E.; et. al.: USAF Stability and Control 
Datcom. Flight Control Division, Air Force Flight Dynamics 
Laboratory, Wright-Patterson Air Force Base, Ohio 45433, 1972. 

32. Engineering Flight-Test Guide for Small Airplanes. Directive 

Number 8110.7, Federal Aviation Administration, Washington, DC, 
1972. 

33. Hamlin, Benson: Flight Testing Conventional and Jet-Propelled 

Airplanes. The MacMillan Company, New York, (1946). 

34. Smetana, Frederick 0.; Summey, Delbert C; Smith, Neill S.; and 

Garden, Ronald K.: Light Aircraft Lift, Drag, and Moment 
Prediction - A Review and Analysis. NASA CR 2523, 1975. 

35. Allen, H. Julian; and Perkins, E. W. : A Study of Effects of 

Viscosity on Flow Over Slender, Inclined Bodies of Revolution. 
. NACA TR 1048, 1951. 

36. Fink, Marvin P.; Shivers, James P.; and Smith, Charles C, Jr.: 

A Wind-Tunnel Investigation, of Static Longitudinal and Lateral 
Characteristics of a Full-Scale Mockup of a Light Twin-Engine 
Airplane. NASA TN D-6238, 1971. 

37. Wenzinger, Carl J.; and Rogallo, Francis M.: Wind-Tunnel Investigation 

of Spoiler Deflector, and Slot Lateral -Control Devices on Wings 
with Full -Span Split and Slotted Flaps. NACA TR 706, 1941. 

38. Gil ruth, R. R. : Requirements for Satisfactory Flying Qualities of 

Airplanes. NACA TR 755, 1943. 

39. Anon.: Detail Specification for Engine, Aircraft, Model L10-CIE6, 

200 Horsepower, direct Drive. Avco Lycoming Specification 
No. 2416. April 13,. 1970. 

40. Pfleegor, Cliff: Determination of Installed Horsepower for 

Lycoming Reciprocating Engines. First Annual General Aviation 
Technology Fest, November 1975. (Oral presentation only.) 

41. Borst, J. v.: A Short Method to Propeller Performance, Curtiss-Wright 

Corporation, Caldwell, NJ, 1956. 

42. Weick, Fred E.: Aircraft Propeller Design. McGraw-Hill Book Company, 
Inc., 1930. 



188 



43. Wood, Donald H.: Tests on Nacelle-Propeller Combinations in 

Various Positions with Reference to Wings. Part I, ThickWing 
NACA Cowled Nacelle-Tractor Propeller. NACA Report 415, 1932. 

44. Wood, Donald H.: Tests on Nacelle-Propeller Combinations in 

Various Positions with Reference to Wings. Part II. Thick Wing 
Various Radial Engine Cowlings-Tractor Propeller. NACA 
Report 436, 1932. 

45. Hoerner, Sighard F.: Aerodynamic Drag. (1951) Published by author. 



189 



APPENDIX A 



PERFORMANCE PARAMETER EXTRACTION METHOD 



WITH ERROR ANALYSIS 



190 



APPENDIX A 

P ERFORMANCE PARAMETER EXTRACTION METHOD 
WITH ERROR ANALYSIS 

This appendix is included for the convenience of the reader as a 
brief description of a method for extracting airplane drag and power 
information from dynamic, maneuvering flight data. The development of 
this method and the preparation of the materials for this appendix were 
done by Dr. Frederick 0. Smetana of North Carolina State University. 
Publication of a full description and the results of flight- test 
applications of the method is planned in a future NASA contractor report. 

MEAS URING DRA G AND THRUS T IN FLIGHT 
The Concept 

Most techniques for the determination of aircraft drag in flight 
rely on the fact that when the aircraft is in unaccelerated flight, the 
forces along its x-axis, principally the thrust and drag, are in balance. 
Then, if one knows the propulsive thrust for a particular flight condition, 
he automatically knows the aircraft drag at that condition. Thus, to 
apply these techniques one must know that V = as well as the propulsive 
thrust as a function of flight speed, altitude, and power setting. 

This, unfortunately, is not determined easily. Although engine output 
can be measured accurately on a test stand as a function of altitude and 
power setting, and propeller characteristics can be determined in a test 
cell as a function of rpm and flight velocity, the flow disturbances caused 



191 



by putting a cowled engine behind a propeller and mounting the whole 
on an airplane are not readily determined apriori. Hence, efforts 
have been made from time to time to measure inflight thrust using such 
techniques as the torque reaction produced by the engine or the vehicle 
acceleration at constant altitude produced by varying power levels. 

The reader will readily appreciate the difficulties which such 
techniques entail. In the case of the ATLIT aircraft, instrumentation 
to measure reaction torques was not available and the longitudinal 
accelerometer provided in the instrument package was not considered a 
primary test instrument, at least initially. Further, the establishment 
of really unaccelerated flight at many different speeds is very consuming 
of flight-test time. It is for these reasons that an effort was made to 
develop an alternate technique to measure thrust and drag simultaneously 
in accelerated flight. 

The origin of the concept is quite simple. Recent workers attempting 
to extract the values of stability derivatives from flight data have all 
faced the problem of fitting an analytical model containing 13 or more 
undetermined coefficients to a set of four or five simultaneous time 
histories. That is, the number of unknowns greatly exceed the number 
of independent equations one can write to describe the motion. The 
problem is usually attacked (see ref. A-l» for example) by fitting the 
equations to the time histories at a number of different times. Theoreti- 
cally, one need only fit the equation the same number of times as one 
desires to find coefficient values. In practice, it is fit many, many 



192 



times and the values which best satisfy the time history in some 
statistical sense are chosen. If the initial estimates of the parameter 
values are reasonably accurate, the procedures usually converge on the 
correct values. However, since the system is not determinant, convergence 
IS not guaranteed. 

The problem in determining both drag and thrust simultaneously in 
flight is that there is one more unknown than there are jequations.. , Mathe- 
matically this means that for any flight condition there are an infinite 
number of sets of T and D which satisfy the equation. For any T- there 
is only one D, but one can find the corresponding D for any arbitrary 
choice of T whether it has any physical meaning or not. 

Following the fairly successful approach used in stability 
derivative extraction, it was reasoned that if one would write the 
equation of motion substituting flight data for different times in the 
flight, he could create a system of equations equal to the number of 
unknowns. Formally, the equation of motion of the vehicle along its 
trajectory in the X-Z terrestrial plane is 

^ + sin Y = ^^^^ (A.l) 

g w 
In order to apply the technique, we wish to express the thrust and drag 
in a polynomial expansion of some easily-measured flight variable with 
the coefficients to be undetermined constants. Now, the thrust is known 
to depend primarily upon flight speed for a given power setting so that 
we choose the representation 



193 



T = ^^ [Pq + PjV + PgV^ . (A. 2) 

In other words, we assume that the power-speed relationship is a parabola. 
Given the characteristics of most propellers, P and P, will be 

positive and P^ negative. We insert the cos a term because we assume 

that the propeller thrust is always applied along the x-body axis rather 
than along the flightpath. Drag, on the other hand, is always defined 
with respect to the flightpath. We can represent the drag by the 
equation 

D = 1/2 pSV^ TCq + C^ a^ + Cj3 a^j , (A. 3) 

where a is measured from zero lift and the sixth power for the third 
term was chosen on the basis of curve fits to some actual data. Note, 
however, that we may alter the model to represent a particular situation 
more accurately without affecting the validity of the procedure. 

Substituting these relationships into the equation of motion yields 



WV J. u c,-„ .. - cos 
— + w sm Y = — y 



^^ \p^ + P^V + P^V^j - 1 pSV^ 



This equation has six unknown but constant coefficients. By determining 
the flight values of y, W, V, V, p, and a at six different times, we 
create a system of six linear equations in six unknowns. This can then 

be solved for the values of P , P., ^o'' ^bii'' ^ff '' ^,"<^' ^^ . 

0.1 1 ^ ; 2 



194 



Difficulties in Concept Execution 

Unfortunately, this system of equations is what mathematicians call 
ill-conditioned; that is, very small changes in any of the measured 
values (a, W, V, V, y» p) can cause the coefficient values (P , P-, etc.) 

to change radically. Further, the solution guarantees to pass through 
the six selected points only. For any other speed, acceleration, angle 
of attack, weight, flightpath angle, or altitude, the thrust or drag 
computed with these six coefficients may be quite wrong. In addition, 
the coefficient values themselves may be ridiculous (for example, a 
negative Cp, value) yet the total drag as determined from 

^n "*■ ^n ct + Cr, 01 may be very reasonable. 
^0 ^1 ^2 

These problems are to some extent traceable to the adequacy of the 

analytical model used. A model which does not well represent what 

actually occurs will, when fit to the data using this procedure, produce 

nonsense numbers for some of the coefficients, i.e., nonsense numbers in 

the physical sense but absolutely correct numbers in the mathematical 

sense. For example, if the speed-power relation should in fact be a 

constant, then an attempt to fit it with a parabola will usually yield 

non-zero values for P, and P„. Thus, a successful solution routine 

must have a provision for examining the results (at least manually) 

for reasonableness and for changing the analytical model if the 

results are not reasonable. 



195 



There is also a problem concerned with the selection of the six 
data sets submitted to the solution routine. The reader will recognize 
that if one selects six points very close together in speed, the data 
must be extremely accurate because all significance can be lost in 
taking the differences between adjacent numbers as one does in solving 
a system of six equations. 

The maneuver selected to generate these data was a pullup-pushover. 
Beginning at the highest speed in a configuration of interest, a pull up 
is initiated and the airplane is decelerated to the minimum speed for 
the test. At that point, a pushover is done, allowing the airplane to 
accelerate back to maximum speed. 

Amelioration of Solution Difficulties 

One means of selecting the six points submitted to the solution 
routine so that it will yield reasonable results is to select those 
points where the velocities are given by 

V, = V . for the maneuver 

1 min 

Vo = V„,„ for the maneuver 

2 max 



^3 = ^1 



V = V 




(A.5) 



196 



^5 = ^4 




^6 = ^5 



[V 



1/5 



This procedure spaces the points over all the available data giving 
emphasis to the portion of the drag curve when changes with speed are 
most rapid. When applied to theoretically-generated data, the original 
coefficients can be recovered to within 1 percent. 

For a variety of reasons, flight measurements will never be as 
accurate or as noise free as theoretically-generated data. One then asks 
the question, "How can I use the remainder of the data (the sets of a, 
P» Y> V, V, W beyond the six sets mentioned above) taken during a 
30-second maneuver to improve the accuracy of the coefficient extraction 
procedure?" The classical answer is fit the assumed form of the curve 
(equation A. 4) to the data by a least-squares technique. What this does 
is to determine those values of the coefficients (P , P., P«, C^ , 



Cp> , Cf. ) which make the sum of the squares of the distance from the 

curve to each of the data points as minimum. 

Data Filtering 

All records of the flight of actual aircraft will contain spurious 
contributions to the data signals arising from electrical noise, 
instrument errors, structural vibrations, and atmospheric turbulence. 



197 



Since the model we have chosen to represent the aircraft does not include 
such effects, it is desirable to remove them, in so far as possible, 
before submitting the data to the coefficient extraction routine. Not 
doing so may cause the extraction routine to produce physically 
meaningless results. 

All filtering schemes proceed from the idea that continuous data 
signals are composites, each signal made up of sine waves of all 
frequencies. Each of these sine waves in the composite has a definite 
amplitude and phase relationship to the other sine waves making up the 
signal. By suppressing those frequencies which, on the basis of analysis 
or experience, cannot arise from the aircraft behavior of interest, one 
can remove most of the spurious contributions to the signal. Tradition- 
ally, filtering was done on continuous signals using frequency-sensitive 
passive networks. In the present case, however, the flight data were 
received in digital form so that the filtering was accomplished 
mathematically using a computer.* 

Since this procedure permits one to describe a signal time history 
in terms of its harmonic content, it is therefore possible to reduce the 
amplitudes of or eliminate certain constituent frequencies from the set 
before regenerating the signal; in essence, filtering out the unwanted 



*The data are, nevertheless, just digitized samples of continuous functions, 
For this reason, we have chosen to employ mathematical techniques more 
appropriate to such functions than the more commonly-used digital filtering 
techniques which seem more appropriate to the analysis of data which are 
inherently trains of impulses. 



198 



contributions to any desired degree, without any disruption of the phase 
relationships among the remaining contributions. This represents a 
level of filter performance far above that possible with passive 
elements in analog circuits. 

Corrections to Measured Accelerations 

The scheme to extract drag and thrust simultaneously from flight 
data has been found to require accurate indications of the acceleration 
along the vehicle's flightpath in order to yield acceptable results. 
Usually it is not possible to locate the measuring instrument (accelero- 
meter) precisely at the vehicle's center of gravity, so it is 
necessary to correct the instrument's indication for this fact and then 
to relate the acceleration along the vehicle's x-body axis to the 
longitudinal acceleration along the flightpath. 

Accel erometers are generally masses constrained to move along the 
axis of a tube and centered by springs at either end. The position of 
the mass relative to the center of a tube is proportional to the 
acceleration and is measured electrically. When the aircraft accelerates 
along the flightpath, the mass moves aft of the center of the tube. Now, 
the same effect is produced when the accelerometer is tilted nose up even 
though there is no acceleration. Thus, it is necessary to subtract a 
term ni sin 9 from the accelerometer indication to account for this 
effect. 

If the accelerometer is located x feet in front of the CG, its 
mass is caused to move forward as a result of the angular rotation of 



199 



Then solving for V, one has 

2 

a -n^ sin e + X q - z q 

V - ind 

cos a 

+ V (a + q) tan a . (A. 12) 

The value given by (A. 12) should now be the sanie as that obtained by 
differentiating the variation of true airspeed with time. 

Corrections to Airspeed Indications 

Of course one does not measure true airspeed directly. An airspeed 
sensor measures only a pressure difference. This difference is affected 
by the sensitivity of the pitot and static pressure sources to angle of 
attack, the disturbance to the free-stream pressure at the static-pressure 
source resulting from the presence of the aircraft, the compressibility 
of the air, and the difference in pneumatic lags of the pitot and 
static-pressure lines. The pneumatic lag also introduces a time delay 
in the airspeed indication. Since the airspeed indicator is calibrated 
for standard sea-level conditions, any variation in atmospheric 
temperature will affect the true; airspeed at a; given pressure difference. 

The theory of the pi tot-static tube assumes that the air is brought 
to rest at the pitot pressure source adiabatically and that the static 
source senses the pressure in the free stream (i.e., away from the 
airplane). With these assumptions, it is easy to show that the true 
flow velocity is related to the measured pressures by 




(A. 13) 



200 



2 
the aircraft by an amount x-q . One must therefore add this term to 

the accelerometer indication. Similarly, if the accelerometer axis 

is located z feet below the x-body axis, then the accelerometer mass 

is displaced rearward by an amount proportional to z-q. 

The linear acceleration along the x-body axis in terms of the 
accelerometer indication location, and angular velocity is therefore 

2 

a„ = a^ -n;sin e + x-q - z-q . (A. 6) 
^ ^ind ^ 

We desire the acceleration not along the x-body axis but rather 

along the flightpath. We know that for motion in the x-z terrestrial 

plane 



a^ = u + q.w (A. 7) 



and 



u = V cos a (A. 8) 

w = -V sin a (A. 9) 

where V is the velocity of the aircraft along its flightpath and u and 
w are components of this velocity along the principal axes of the 
aircraft. In terms of (A. 8) and (A. 9) 

a = V cos a - V ct sin a - q V sin a 

= V cos a - V (ci + q) sin a . (A. 10) 

Equating (A. 6) and (A. 10) yields 

2 
3^ - n^ sin e+q -zq=V cos a - V (d + q) sin a 

^ind 

(A.ll) 



201 



where P 1s the altitude pressure, q is the difference between the 

pitot and static pressures, T is the local free-stream absolute 
temperature, R is the gas constant for air and y is the ratio of 
specific heats of air (1.4 for diatomic gases at normal temperatures). 
The P indication for use in this equation comes from the static-pressure 
source of the pitot-static tube and the T indication from a temperature 
measuring device. Since one cannot measure the local free-stream 
temperature readily while the vehicle is in motion, temperature sensing 
devices most often measure the stagnation temperature, T , which is 

related to the free-stream temperature by 

T = ^^ (A. 14) 



- 



LP 
In terms of the stagnation temperature, the true airspeed is given by 




Unfortunately, it is usually not possible to locate the static-pressure 
source on an airplane in a region where the static pressure is the same 
as the free-stream value. Hence, the static-pressure indication is in 
error by an amount AP. This "position error" so called is felt in both 
the altitude and q indications. If we call P' the measured altitude 

pressure and q the measured pressure difference, then because 

q^' + P- = q^ + P = P^ , (A.16) 



202 



and 



P = P' - AP 



one can write 












%*" 


_^c' 


+ P' 


1 


+ 


p* 


p 


P' 


- AP 


P' 


- 


AP 



(A. 17) 



(A. 18) 



in terms of the measured values and the static-source position error which 
is usually determined by flight calibration and is expressed in terms of 



AP 



as a function of q ' or indicated airspeed. With this effect 



included the expression for true airspeed becomes 




2YRTg 



1 - 



1 + 



AP 

^c 




(A.19) 



Fortunately, modern pitot-static tubes are relatively insensitive 
to changes in angle of attack so that the q ' and P' indications do 

not depend on the tube's inclination to the airstream over the useful 
range of aircraft angles of attack. The position error, however, does 
depend upon angle of attack and aircraft configurations. At steady 
speed and constant weight the position error can be related, as it 
commonly is, to q ' or indicated airspeed, but during maneuvers it may 

be necessary to employ a correlation with angle of attack instead. 
Whether this is necessary may be determined by calibration. If it is, 
one must then determine true airspeed and true angle of attack iteratively. 



203 



The compressibility correction mentioned earlier is already included 
In (A. 19). Conventional low-speed acirspfeed indicatons, it may be noted, are 
simply mechanizations of the equation 

V. = -— ^ , (A. 20) 

where p is the mass density of the air at standard sea-level conditions. 

If the airspeed indicator calibration includes compressibility effects, 
equation (A. 13) with standard sea-level pressure and temperature is 
mechanized. 

Effect of Pneumatic Lag on Dynamic Airspeed I nd icjti pns 

If pneumatic signals transmitted through the pitot and static lines 
travel at different speeds then the q ' and P' values will be in error. 

In most aircraft with pressure sensors located in cabin area, the 
pneumatic lines are long enough that their response characteristics can 
be considered analogies to those of resistance-capacitance electrical 
circuits. The resistance is proportional to length/(diameter) while 
the "capacitance" is proportional to system volume. Since the static 
system includes more instruments than the pitot system and, frequently, 
larger volumes, the static-line diameter must be larger than the pitot 
or a restriction must be placed in the pitot line in order to keep the 
response times equal. Even if the line responses are equal, V' 

will lag the correct value by a time which is proportional to h and V. 
Corrections for these lag effects will be necessary if the time constants 
of the pneumatic lines in ATLIT exceed about 0.1 seconds. 



204 



The lag corrections are applied as follows: 
let 

P„ = the instrument indication 

Pj = correct pressure 

Assume Poiseui lie-type flow where the mass flow is proportional to the 
pressure difference to the first power: M ~ aP. The rate of change of 
pressure in a volume connected by long tubing to the atmosphere is, for 
isothermal conditions, simply 

P2 = 7 (Pi - ^2^ (A. 21) 

where x is an experimentally determined time constant- In the case 
where P, is changed instantaneously and held at the new value, one may 

write 



^^2 dt 



(A. 22) 



or 



(Pj - P2^ ^ 



-In (P^ - Pg) = t/T + C 

P, - P2 . ce-^/^ <A-23) 

when t = 0, Pp = Pp 



when t ->■ «>, P^ -> p (A. 24) 



205 



thus 



P^ - P^ e"*/'' + Pj (1 - e"^/"^') . (A.25) 

^ 



This says that by differentiating P (t) to obtain P at any ti 



2 ^ ' "*" '2 



me 



we can find P^ by taking P^ and adding tP^ (t). 
If we now call 



P the static pressure 



P^ the stagnation pressure 

T^ the time constant in the static system, seconds 

T^ the time constant in the stagnation system, seconds 



then the lag-free value of q ' is given by 



m m ""m m 

= ^c' -^-^2 V - T, P^ (A.26) 

m m s^ 

and the lag-free value of P is given by 

Ps " f's *" ^1 ^ (A. 27) 

m m 

where the subscript m indicates the recorded value. Where x, = t«, 

1 2* 

(A.26) can be written 



•c -"c^ "^"c . (A.28) 



^ m ^ 
This value of q^ , corrected for lag, is then applied in equation (A. 19) 



206 



Correction of Angle of Attack Indications 

In addition to factors such as transducer linearity, gain, and bias, 
the angle-of-attack indication is affected by the presence of the 
carrying aircraft and by its rotation. It will be recognized that for 
an angle-of-attack vane located x feet ahead of the e.g. an incremental 
angle 

Aa = tan"-^ ^ . (A. 29) 

must be subtracted to the transducer indication to account for vehicle 
rotation. In addition, there is usually a relationship of the type 

^'true = S ^'indicated "" ^2 ' ^^'^^^ 

between the angle of attack measured in the neighborhood of the aircraft 
and the true (i.e., at infinity) angle of attack. The values of C^ and 

C^ depend upon the location of the vane relative to the aircraft and the 

geometry of the aircraft. They are therefore almost always found from 
flight calibration tests, since the flow field about a complex, shape, such as 
a' complete aircraft is almost impossible to determine analytically. 
Assuming that these coefficients are known, one may write the expression 
for true angle of attack as 



"true ' ^ 



1 {"indicated - t^""^ H}^ Cg . (A.31) 



Note that the value of V used in (A, 31) should be that obtained from 
(A. 19). One may then smooth o'+^ye^^^ ^^^ compute the derivative, 
a(t), by the Fourier analysis procedure. 



207 



Determination of p(t) 

Equation (A. 4) requires as an input p(t). 
determined from 



= (P' - AP) 



This is readily 







y-1 


1 + 


p' 


Y 


P' 


AP 




L^c 


\'j 





(A. 32) 



If the altitude pressure transducer is calibrated in feet, then the 
appropriate pressure versus altitude function must be employed to convert 
the indications to pressure values. 

Conditioning of Other Data Inputs to the Drag 
and Power Extraction Method 

In addition to the velocity, angle of attack, and atmospheric 
density, W(t) and 6 are required as inputs. Fortunately, for the 
maneuvers of interest W changes so little that it can be taken to be 
constant or, at most, varying linearly during a maneuver. Usually 9 
requires no corrections beyond the instrument calibration if the erection 
mechanism i-s disabled during the maneuver. Since the indication is 
sampled and since there may be electrical, airframe, and turbulence- 
induced noise, smoothing may still be necessary. This is also true for 
the pitch- rate indication, q, which is used in the C, computation 

and the a and a corrections. 

More General Power and Drag Models 

In a normally aspirated engine, the manifold pressure and, hence, 
the power output for a given throttle setting will usually vary directly 



208 



with the atmospheric density. Thus, if the maneuver to provide data 
for the power and drag extraction process involves a change in altitude, 
there will be a change in power at a given speed corresponding to the 
change in p even if the pilot does not change his throttle setting or 
rpm. To account for this, we need to multiply the expression for power 
by Bairstow's equation (ref. A. 2) 

Pref/Po ■ °-^^^ 

-^^ — (A. 33) 

p/p^ - 0.165 

where p is the standard sea-level value of p and p ^ is the value 

of p at the beginning of the maneuver. 

The parabolic form of the speed-power relation used in equation (A. 4) 
is obviously satisfactory over small differences in speed and should 
represent the thrust horsepower of fixed-pitch propellers reasonably well 
over most of the aircraft's speed envelope. The higher efficiency levels 
provided by a constant speed propeller at the lower speeds, however, 
makes it necessary to employ a higher order polynomial or other function 
having additional degrees of freedom (coefficients) to represent the thrust 
horsepower adequately over a wide speed range. Variants of one such 
function were chosen for further study: 

p 

P = P, + —^ + P,V + P.V^ + P^V"^ (A. 34) 

^ yl/2 J ^ t> 



209 



These variants include 



P = Pj (A. 35) 

P = P^ + PgV (A. 36) 

P = P, + Zl_ (A. 37) 

P = P^ + P^V + P^V^ (A. 38) 



P = P^ + -^ + P3V + P^V'^ (A. 40) 

P = P^ + P3V + P^V^ + PgV^ (A. 41) 

One will note also that the drag expression is really satisfactory 
only if a is measured from zero lift. Since the angle reference for 
flight data is quite arbitrary, it is difficult to establish the angle 
for zero lift apriori. To accommodate an arbitrary reference, i.e., to 
replace a by a in equation (A. 3), requires that the representation 

for Cpj contain all powers of a through 6. We choose, however, to 

investigate only variants of the following form: 



210 



These variants are 

The three drag expressions and the eight power expressions give us 
a total of 24 analytical models with which we can attempt to fit 
experimental data. It will probably be necessary to employ all of the 
models or at least this numberof models until experience with data for 
a particular aircraft permits one to discard these models which do not 
apply. 

One may also ask why should one also employ a model which is simply 
a reduced form of a more general model? The answer lies in the extreme 
sensitivity of the coefficient solutions .to small errors in the data. 
Generally, the more general models are more sensitive to these errors 
so that under these circumstances a simpler form may yield reasonable 
results, whereas the more general form may yield nonsense numbers. It 
should be recalled that since any power, if accompanied by a suitable 
drag, will solve the equation of motion, these physically absurd numbers 
are legitimate mathematical solutions. How then does one determine 
whether the solutions obtained are reasonable? 

The first means of assessing the reasonableness of the solution set 
is to use them along with the experimental data in the proper form of 
equation (A. 4) to compute an error term, S. 



211 



N 
S - E 
i=l 



W V. 

_!_L = W^. sin (0j - a^) P^ - cos a^ P^ 



U9 



cw 2 2 

P-iSV. p.SV. 2 
V. COS a^ P^ + ^^ C„^ + !l_l_ C.2 Cn^ + 

a- C 



(A. 45) 



2 i Dp 

-13 
For 300 data points, a value of S < 10 generally indicates 

coefficient values within 1 percent or so of the correct values. (For 

-21 
the exact coefficient values, S < 10 .) Coefficient values in error 

by 5 percent, for example, may still be of interest but with errors of 

this size it may become difficult to identify the best model and 

coefficient set merely by checking to see which model gives the smallest 

value of S. S . will now be on the order of lO' for 300 points, but 

the coefficient set for S • may give absurd powers and drags. For 

this reason,- it is desirable to add a second constraint which an 
acceptable model and coefficient set must satisfy: The horsepower for 
any speed must be positive and less than Y (Y = 400 for ATLIT); C-. 

must be positive and less than Z (2 = 0.12 for ATLIT) for any a. 

One frequently finds that with noisy data yery few of the 24 coefficient 

sets satisfy this second constraint. 



212 



Ef fect of Data E^rrors on Coefficient Extractions 

We have noted above that by operating on exact data, it is possible 
for the coefficient extraction procedure to recover the values of the 
coefficients in the power and drag polynomials to six significant figures 
We have also noted that this procedure is quite sensitive to data 
inaccuracies. In order to place some quantitative bound on this 
sensitivity, the exact input data were artificially degraded and resub- 
mitted to the coefficient extraction procedure to determine how the 
coefficient values were altered. Two types of degradation were employed: 
random noise and constant bias. For the random noise, a random number 
generator was employed at each 0.1 seconds of each trace and the output 
scaled so as to be 1 percent of the maximum value of the function, e.g., 
1 percent of the maximum value of V(t) during the maneuver. These 
scaled noise values were then added to the exact function values to 
obtain the degraded data. For this experiment, all data which would 
normally be measured in flight were degraded. This was too noisy. No 
coefficient set would satisfy the reasonableness criterion. 

In an experiment, the data traces were degraded individually by a 
constant bias error. Reproduced as figures A.l through A. 10 are the 
recovered speed-power and drag-alpha characteristics for various bias 
errors compared with the undegraded characteristics used to generate 
the data traces. Generally, the characteristics for the largest bias 
error which can yield reasonable results are shown along with the 
characteristics for amaller bias errors so that the reader may assess 
the linearity of the change in characteristics with the change in bias 



213 



error. Note that weight and altitude bias errors of the magnitude shown 
are not particularly serious. As might be expected, bias errors in 
airspeed affect the power determination primarily and have little influence 
on drag. The same is true with regard to bias errors in V. Bias errors 
in e and a, however, are extremely destructive. Even a 0.7 
error in 6 results in about a 10-percent error in C^ while a -1.9 



error in 9 results in an error of about 37 percent in Cp . The case 



for a bias error of +1.9 failed (e.g., gave a power exceeding the limit 
of 400 hp). An angle-of-attack bias error of as little as 0.1 is 
noticeable in the final result, while an a bias error of 1.6 results 
in drag and power errors in the neighborhood of 30-40 percent. In 
addition, the shapes of the curves are altered drastically. 

These results demonstrate the extreme sensitivity of the coefficient 
extraction procedure to typical noise and instrument errors encountered 
in flight- test work. This is true even after the data have been filtered 
to remove the noise components which occur at frequencies above the usual 
aircraft responses to control deflections. Thus, to obtain accurate 
drag and power data using this procedure some means must be employed to 
reduce the noise components in the data at what might be termed signal 
frequencies- 

Reduction of Noise at Signal Frequencies 

The filtering technique discussed previously has been shown to be 
highly effective at suppressing noise components in the data at 
frequencies above the principal components of the aircraft response. At 



214 



frequencies below this cutoff value it is impossible to separate the 
noise from the signal without resort to additional information! The 
only additional information universally available are the relationships 
among the measured variables and the power and drag as functions of 
time, i.e., the equations of motion and their auxiliary equations: 



V = 5^ -tf-S f^J P^^> -9^^'"^ 



Y = ^ f- Cj_ (a) P(h) - ^ COSY + ^PsHia 



Y 



W = -cP 
Fi = V siny 
X = V cosy 



p(h) = pQ (1 - 6.86 X 10"^ h)"^*^^ 
a (a=o) 



Cj^(a) = Cjj + k CL(a) 2 + k^ Cj_(a) ^ 



P < P (h, V) 
— max ^ ' 



A. 46) 

A. 47) 

A. 48) 
A. 49) 
A. 50) 

A. 51) 
A. 52) 

A. 53) 
A. 54) 



This system of equations can be solved simultaneously to yield the 

variable values as functions of time provided any two of these (a, 9, V, 

h, W, P) are known apriori. C^j , k, k,, kp, C, , and C, must also 

a=0 a 

be assumed known in order to carry out this procedure. Then, given 

values for these constants and W(t) and h (t), one can solve the system 

a self-consistent set of a(t), 6(t), V(t), and P(t). Since these time 



215 



histories are related in a consistent fashion through the equations of 

motion and since noise will likely affect each time history differently, 

it is then possible to identify the noise present in each time history 

at signal frequencies by comparing the solutions to the measured data. 

Once the noise is identified, one can take steps to reduce or remove it. 

Unfortunately, one does not have apriori a very accurate indication 

of Cpj , k, k,, kp, C. , and C. so that one's knowledge of these 
a a=o 

coefficient values will improve in the process. Convergence to the 
correct values cannot be guaranteed. 

Lift Computation 

Once the power into the airstream has been determined, it becomes 
a relatively straightforward task to determine the lift time history. 
We note that the equation of motion of the vehicle in the direction 
normal to flightpath in the terrestrial x-z plane is 

L - W COSY + T sina = Vy W (A. 55) 

g 

from which we may easily obtain 
2W 



[| (^ - ^) - "IJr^ ■" cos(e - a)] . (A. 56) 



^^ ^ PSV2 I? 

Presumbably, W, p, V, 6, 0, and a are measured directly as functions of 
time during flight while a is obtained by differentiating a(t). A value 
for P is also a result of the process which extracts Cpj(c(). This P 

may be stated as a function of V alone (as in equation (A. 4)) or as 
a more general function, say in terms of V and p. In either case 



216 



substitution of this value in the foregoing equation then yields C, (t). 
By correlating C, (t) with a(t) it is possible to develop C, (a) 
as v;ell as C, /C^ as a function of a. 



217 



APPENDIX A 



REFERENCES 



A.l Taylor, Lawrence, W. : A System Identification Using a Modified 
Newton-Raphson Method -- A Fortran Program. NASA TN D-6734, 
1972. 

A. 2 Oswald, W. Bailey: General Formulas and Charts for the Calculation 
of Airplane Performance. NACA TR-408, 1932. 



218 





300 


pf 




275 


- 




250 


- 




225 


' 


£ 


200 


- : 


n 


175 
150 




1 

a. 


125 


- 








100 


■ 




75 


- 




50 


- 




25 



■V 




O.ZQ 



Bias: Weight + 444. S N (100 1bf) 



100 no 120 130 140 150 160 170 



Airspeed V, knots 



(a) Extracted power coefficient 



"I I I I I I r 



0.18 - 
0.16 - 
0.14 - 




J I 1 L 



1.0 2.0 3.0 4.0 5-0 6.0 7.0 8.0 9.0 

Angle of attack o, deg 

(b) Extracted drag coefficient 
Figure A.l- The effect of weight biasing. 



219 



300 HVf- 

Z75 



250 
225 
20Q 

X 

£ 175 

°S 150 

I 125 

100 

75 

50 

25 I- 



Bias: Altitude + 152. 4 ni (500 ft) 



120 130 



Airspeed V, knots 

(a) Extracted power coefficient 



200 S 




1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 
Angle of attack a, deg 

(b) Extracted drag coefficient 
Figure A. 2- The effect of altitude biasing. 



220 



300 
275 
250 
225 
200 
175 
150 
125 
100 



r^ 



Bias: Airspeed + 0.1805 m/s (0.5921 fps) 



go 



J I \ I I I L 



100 110 



120 130 WO 

A'Irspeed V, knots 



150 160 170 



(a) Extracted power coefficient 



1 r 



n r 



J I L 



Bias: Airspeed + 0.1805 m/s (0.5921 fps) 



J 1 L 



1.0 2.0 3.0 4.0 5,0 6.0 7,0 8.0 9.0 10.0 



Angle of attack a, deg 



(b) Extracted drag coefficient 



Figure A. 3- The effect of airspeed biasing 



221 



300 

275 - 

250 - 

225 - 
200 
175 
150 

125 - 

100 - 

75 - 

50 - 

25 - 



"1 r 



I I 




Ury. 



0.20 
0.18 
0.15 
0,14 
0.12 
0.10 
0.08 
0.06 
0.O4 
0.02 



Bias: Airspeed + 0.914 m/s (3 fps) 



J I L 



1 1 I 



Bias: Airspeed + 0.914 m/s (3 fps) 



Assumed 



250 t 



100 110 120 130 140 150 160 170 



Airspeed V, knots 

(a) Extracted power coefficient 




) J_- _! _J 1 I I 



1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 
Angia of attack a, deg 

(b) Extracted drag coefficient 
Figure A. 4- The effect of airspeed biasing 



222 



300 

275 h 
250 
225 
200 

^ 150 

100 
75 - 
50 - 

25 - 




"T I 



^-^r^ 



Bias: Acceleration + 0.0305 m/sec (0.1 ft/sec ) 



J I I I I L 



100 no 120 130 140 150 160 170 



Airspeed V, knots 



(a) Extracted power coefficient 



0.20 



0.18 - 

0.16 

0.14 

0.12 

0.10 

0.08 



1 r 



-| 1 1 1 1 I r 



Bias: Acceleration + 0.0305 m/sec (0.1 ft/sec L 



J I I L 



J L 



1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 



Angle of attack a, deg 



(b) Extracted drag coefficient 
Figure A. 5- The effect of acceleration biasing 



223 



300 
Z7S 
250 
Z25 
200 
. 175 
I 150 

■0 

"■ 100 
75 




Bias: Acceleration + 0.3048 m/sec (1.0 ft/sec ) 



o'A^ 



lOD no 120 130 140 150 160 170 



Airspeed V, knots 



(a) Extracted power coefficient 



0.20 
0.18 


- 


1 


1 


1 1 


1 1 1 - 1 1 


- 


0.16 


- 










- 


0.14 


- 












0.12 


- 










/ 


0.10 


- 






Bias: 


Acceleration + 0.3048 m/sec^ (1.0 ft/sec^l/ 


/ 


0.08 


- 








^/^ 


- 


0.05 


- 






Biased > 


. ^^^^^-^'^"^ 


- 


0.04 


_ 






Z""^ 




- 


0.02 


- 




Assumed ^ 






1 


\ 


1 1 


.1.1 1 1 1 





1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 
Angle of attack a, deg 

(b) Extracted drag coefficient 
Figure A. 6- The effect of acceleration biasing 



224 



300 
275 
250 
225 
200 
175 

tso 

325 
IQO 
75 



0^- 



"I i i r 



"1 r 



Bias: Pitch angle + 0.7 degree 



J I I 



J L 



lOO no 120 130 140 150 160 
Airspeed /, knots 

(a) Extracted power coefficient 




1-0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 



Angle of attack o, deg 

(b) Extracted drag coefficient 



Figure A. 7- The effect of pitch angle biasi 



ng 



225 



300 
275 
250 
225 
200 
175 
150 
125 
100 
75 
50 
25 



rvr 



T r 




w^ 



0.20 
0.18 
0.16 
0-14 
0.12 
0.10 
0.08 
0.06 



ias: Pitch angle - 1.9 degrees 



no 120 130 140 



Airspeed V, knots 



150 160 



(a) Extracted power coefficient 



1 \ I r 




I ■■ .^ I 



1.0 2.0 3.0 4.0 5.0 6.0 7.0 B.O 9.0 
Angle of attack a, deg 

(b) Extracted drag coefficient 
Figure A. 8- The effect of pitch angle biasing 



226 



rvi 



300 
275 
250 

225 

200 - 

175 - 

150 

125 - 

100 
75 1- 
50 - 
25 



"i i i i I 1 r 



,^A^ 



0.20 
0.18 
0.16 



0.10 
0.08 
0.O6 
0.04 



Bias: Angle of attack + 0.1 degree 



90 100 110 120 130 



Airspeed V, knots 



(a) Extracted power coefficient 



I - -\ 1 • — 1 r 



J \ I I \ L 



150 160 170 




1.0 2,0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10,0 
Angle of attack a, deg 

(b) Extracted drag coefficient 
Figure A. 9- The effect of angle of attack biasing 



227 



1 r 




,LaA 



100 no 120 130 



150 160 



Airspeed V, knots 



(a) Extracted power coefficient 



0.16 
0.14 
O.IZ 
0.10 
0.08 
0.06 
0.01 



Bias: Angle of attack + 1.6 degrees 




J I I 1 ! L 



1.0 2.0 3.0 4.0 5.0 5.0 7.0 8.0 9.0 10.0 

Angle of attatfc a, deg 

(b) Extracted drag coefficient 
Figure A. 10- The effect of angle of attack biasing, 



228 



If 



APPENDIX B 

POSITION-ERROR CALIBRATIONS FOR 
STATIC PRESSURE AND ANGLE OF ATTACK 



229 



APPENDIX B 

POSITION-ERROR CALIBRATIONS FOR 
STATIC PRESSURE AND ANGLE OF ATTACK 

The purpose of this appendix is to present instrumentation, methods, 
data reduction, results, and conclusions on position-error calibrations 
for static pressure and angle-of-attack measurements. It should be 
noted that the pressure calibrations are performed only for static 
pressure; that is, it is assumed that total pressure is measured accurately. 

STATIC-PRESSURE POSITION-ERROR CALIBRATION 
Instrumentation and Equipment 

In addition to the onboard instrumentation described in Chapter 4.2, 
the static-pressure calibrations required the following additional 
equipment. 

Trailing Anemometer Installation .- The installation of the trailing 
anemometer on ATLIT is shown in figure B.l. The self-contained system 
consists of: (1) an anemometer airspeed sensor which is trailed from 
the aircraft by a cable in the undisturbed airflow, (2) a mechanism to 
deploy and retract the cable which supports the sensor, (3) the operator's 
control box, and (4) a 27 volt d.c. power supply to drive the deployment 
mechanism. Details on the installation and operation of this device are 
found in reference B.l. Details on the design and construction of the 
device are presently unpublished. 



230 



Based on wind-tunnel calibrations, the anemometer airspeed sensor 
is accurate to within +0.5 kt of true airspeed. With data system accuracies, 
this degrades to ±l.kts. The computed location (reference B.2) of the 
device below and behind the aircraft is shown in figure 8.3. These 
locations are based on an extension of 30.48m (100 ft - 2^5 wingspans) of 
cable. Also, based on calculations of reference B,2, the anemometer 
locations shown in figure B.2 will result in an airplane-induced error 
in measured airspeed less than 0.23 percent. This airplane-induced error 
converts to an airspeed-position error of less than 0.3 kt (0.35 mph) 
at 130 kts (150 mph). 

Tower Fly by Equipment .- The tower flybys were performed at NASA 
Wallops Flight Center. The only airplane equipment required for the 
method was a C-band transponder beacon compatible with the AN/FPQ-6 
radar at Wallops. The AN/FPQ-6 radar facility was used to produce a 
time history of the aircraft location during the test runs. The angular 
precision of this radar is ±0.05 mils (RMS) (from unpublished data). The 
tower involved in the flyby maneuvers is located 3287.6 m (10,786 ft) from 
the radar antenna (see figure B.3). At that range, the angular precision 
gives altitude within ±0.2m (±0.5 ft). The effect of this amount of 
altitude error on the airspeed calibration parameter, Ap/q' , is shown 

in figure B.4. An altitude error of ±0.2m (±0.5 ft) results in a maximum 
airspeed error of ±0.4 kts (at 50 KCAS). This error diminishes with 
increasing airspeed. 

A 76.1m (250 ft) meteorological tower was used for the flybys. The 
tower was equipped with barometric, wind, humidity, and temperature 



231 



recording devices at the 45.7, 61.0, and 76.2 m levels (150, 200, and 
250 feet, respectively). Tower barometric pressure measurements were 
accurate to ±16.76 Pa (±0.35 psf), and tower temperature measurements 
were accurate to ±0.6^ C (±1*^ F). The effects of the temperature errors on 
the meteorological tower are insignificant. 

Table B.l presents a summary of the accuracies involved in determining 
velocity by either the trailing anemometer or the tower flyby method. 
Sample velocity errors have been computed at the velocity limits for each 
method to illustrate the approximate magnitudes of overall accuracies. 
The errors indicated for flight-measured variables are derived from the 
instrumentation accuracies listed in Chapter 4-3. The table shows that 
with the trailing anemometer method airspeed can be calibrated within 
±1.4 knots (one standard deviation) at the low-speed end, and within 
±0.9 knots (one standard deviation) near the upper speed limit for this 
device. Accuracies for the tower-flyby method range from ±2.6 knots 
(one standard deviation) at 90 knots, to ±1.3 knots (one standard deviation) 
at about 170 knots. 

This analysis of accuracies makes it evident that the trailing 
anemometer, not considering its airspeed limitations, is more accurate 
than the tower-flyby technique. Also, the tower flyby requires much 
greater accuracy for flight-measured static pressure. A 1500m (5000 ft) 
altimeter was used on ATLIT for this purpose; therefore, the accuracy in 
flight-measured p is better for the tower flyby than for the trailing 
anemometer where a to 102.4 kPa (15 psia) pressure transducer was used. 



232 



A large error can be tolerated in flight-measured p with the trailing 
anemometer method, but not with the tower-flyby method. 

Experimental Met hods and Flight-Test Programs 
Static Pressure Calibrati on with the Trailing Anemometer.- Two 
calibration techniques were used with the trailing anemometer device. 
First, steady-state data were gathered in the conventional manner, during 
unaccelerated level flight. Second, quasi-steady-state data were gathered 
during slow decelerations (a less than one knot per second) in level flight. 
The deceleration is accomplished by gradually reducing power starting with 
power required for the maximum speed in the configuration of interest and 
bleeiding power off until the stall occurs. The power-off data were gathered 
by idling the throttles at the maximum speed of interest in a configuration 
and decelerating to the stall. 

With either the static or the continuous deceleration method (steady 
state or quasi-steady state, respectively), the calibration theory is the 
same. At each speed of interest during a test run, the true airspeed from 
the trailing anemometer can be compared to the true airspeed as computed 
from the onboard measurements of dynamic pressure, static pressure, and 
temperature. The difference in velocities is the position error. This 
position error will be presented as static-pressure error. 

No demanding pilot techniques are required for either the static or 
dynamic trailing anemometer methods. During the continuous calibration 
maneuver, a simple form of quality control is accomplished by timing the 
maneuver from beginning to end to determine that the average flightpath 
deceleration is less than one knot per second. This assures the effects 
of pi tot-static-system pneumatic lag and time-dependent aerodynamics can 
be ignored. It is also necessary that the throttle (s) be smoothly 



233 



retarded with as little "jockeying" as possible. Jerky motions during 
retarding of the throttles can result in fore and aft swinging of the 
trailing anemometer, making data reduction difficult. 

Static Pressure Calibration by Tower Flyby .- The procedure for 
static-pressure system calibration by tower flyby consists of flying the 
airplane at the same geometric altitude as a fixed-barometric pressure 
recording device. The static-pressure error is determined by comparing 
the static pressured measured onboard the airplane to the pressure 
measured with the fixed-barometric pressure recording device on the tower. 

The test-pilot technique for the tower flyby consists of passing the 
tower at constant power setting while striving to maintain constant 
altitude. During these constant-power, constant-altitude flybys, airspeed 
is allowed to vary. Of the two, airspeed and altitude, it can be shown 
that accurate determination of altitude at the tower passage is critical 
to the overall accuracy of this method. 

Flight-Test Programs 

The methods used for the calibration of the static pressure measuring 
system on the ATLIT were determined by the equipment readily available for 
the task. The use of two overlapping methods, trailing anemometer and 
tower flyby was necessitated by the limitations of each. In general, the 
trailing anemometer covered the low-speed end of the flight envelope and 
the tower-flyby method covered the high-speed end. 

The tests were conducted in smooth air. Table B.2 presents the 
configuration/airspeed combinations for which calibration tests were 
performed. 



234 



Test Conditions for the Trailing Anemometer .- Calibration tests with 
the trailing anemometer are limited to a maximum speed of 165 knots 
(190 mph). This is the speed at which cable instability is predicted for 
the anemometer device used. An additional limiting consideratiion is the 
maximum cable trail-back angle which is considered safe for the airplane 
on which the device is installed. The trail -back angle for the 
installation on ATLIT was computed at 135 knots to be about 
5 degrees (from horizontal, at the aircraft), which allowed for 
safe clearance between the cable and the airplane empennage. 

Tests were run to determine the effects of landing gear position, 
flap deflection, and power on static-pressure position error. The testing 
was done at a pressure altitude of about 305m (1000 ft). The airplane 
weight during these tests varied from 17570 N (3950 lb) to 17348 N (3900 lb) 
at a CG of about 16-percent mac. If position-error data are plotted against 
angle of attack, neither weight nor CG location will have a significant 
effect on the static-pressure error at the noseboom. However, since 
position error as a function of airspeed (V ') is more readily interpreted, 
the present data appear plotted in this manner with V^' corrected to 
gross weight (4200 lb). 

Test Conditions for the Tower Flybys .- Calibration of static-pressure 
error by the tower-flyby technique is limited by safety considerations to 
speeds above a certain minimum. A safe margin above stall speed is 
required because of the close proximity of the airplane to the ground during 
the passes by the meteorological tower. For ATLIT, this meant tower flybys 
could be done at speeds as low as 85 knots, or a speed margin of about 



235 



25 percent above stall speed (flaps 0°). In addition, for speeds below 
85 knots (flaps 0°), it becomes difficult to maintain the required 
levelj, constant-altitude flight past the tower. 

All tower-flyby tests were conducted with flaps up. The airplane 
weight during these tests varied from 17615 N (3960 lb) to 17259 N 
(3880 lb) at a CG of about 14-percent mac. 

DATA REDUCTION 

Trailing Anemometer Data Reduction 

Data reduction methods are the same for both the static and the 

dynamic trailing anemometer techniques. 

\ 
The static- pressure error is defined as 

\ 
Ap = p'-p (B.l) 

where p' is measured onboard the aircraft and p is the ambient static 

pressure. For the speed range of the present tests, incompressible flow 

can be assumed 

qc = q = 1/2 p V^ (B.2) 

where 



P = P'-Ap 
RT 



(B.3) 



Also 



qc = Pt"P " ^'c ■*■ ^P • ^^-^^ 



236 



Equating (2) and (4), and simplifying yields an equation for 
static-pressure (position) error 



^ 



;2 



Ap = ^"^ ' q'c (B.5) 

^ ■*■ 2RT 

This derivation appears in detail in reference B.l. In equation (B.5), 
V is measured with the trailing anemometer and p' and q' are 

measured with the noseboom. Temperature, T, is corrected for adiabatic 
temperature rise based on measured temperature, T. 

T = T, ^ (B.6) 

1 + ^ e M^ 

where e = 1.0 for the ATLIT temperature probe. 

In order to handle high sample rate data (10 samples per second) 
from the continuous calibration maneuvers, the data reduction method was 
programed for a high-speed digital computer. The program expedites 
averaging and smoothing the data over selected time intervals of the test 
run. The resulting data may be either manually faired or numerically 
curve fitted. Data presented in-the figures of this appandix -have been 
manually faired. 

Tower Flyby Data Reduction 
In the tower-flyby test runs, the difference between the height of 
the airplane and the height of the barographic device in the tower 
averaged about 4m (about 13 ft). Therefore, a standard lapse-rate 



237 



correction 1s applied to the tower-reduced static pressure by a form of 
the hydrostatic equation 

-Az 
'c ~ ^l '' 



Ap, = Pi (e "^"^l -1) . (B.7) 



The actual lapse rate for this correction may be computed based on data 

from different sensors on the tower. 

Then 

p = Pj + Ap^ . (B.8) 

The true static pressure, p, is thus determined at the airplane 
geometric altitude by the standard lapse rate. Pressure, p., and 

emperature, T,, are measured on the meteorological tower using values 

from the barograph which is closest to the airplane at tower passage. 

The difference in geometric altitude between the aircraft and the barograph 

closest to the aircraft is Az 

AZ = z^ - z^^ . (B.9) 

Once the atmospheric-static pressure (p) at the airplane geometric 
height has been determined, the static-pressure error is the difference 
between p and the static pressure measured with the airplane noseboom 
(p') at the time of tower passage 

Ap = p' - p . (B.IO) 

RESULTS AND DISCUSSION 
A comparison is presented in figure B.5 of static and continuous 
trailing anemometer data and tower-flyby data (flaps up). It is shown 
that data gathered by these techniques fall within the same region of 



238 



experimental scatter. The bars on the trailing anemometer static-run 

points indicate the extremes in calcula ted re sults due to instrumentation 

errors. 

At cruise speed for ATLIT (a == 0), -^ ~ 0.015. Data. (3) predict 

c 

X 
that a boom (^ = 1.0) on a conical body of revolution (nose shape) 

yields ^ = 0.01, and on a parabolic body of revolution, yields 
^c 

^ = 0.04 (both at M = 0.21, a = with no lifting wing body only), 
^c 

This agreement between predicted and flight- test values of -?— is 

^ c 

explained by the shape of the ATLIT fuselage nose resembling some combination 

of the parabolic and conical nose shapes tested in reference B.3. No data 

exist which allow accurate prediction of — ?— for a given airplane 

^ c 

configuration with varying a, M, and 6^. The shape of the ATLIT 

airspeed-calibration curve agrees with trends for noseboom installation 
in reference B.3. ""' 

Figure B.6 presents the effects on flaps-up airspeed calibration of 
power on and off and landing gear up and down. It was found that the 
gear effects were minimal; therefore, all data are presented gear up. 

Figure B.7 presents the effects of increasing flap deflections on 
airspeed-calibration curves. It is noted that the shape of the curve is 
largely unaffected by flap position, but the location of the curve shifts 
with changing flap deflection. The effect of power-on position error is 
greater with flaps deflected 30 degrees than with flaps up. 



239 



In general, the reduction in position error ^2_ achieved by 

mounted pressure sources on a boom is smaller in magnitude than the 
accuracy with which these position errors may be calibrated. Thus, the 
error in a calibrated boom-mounted pressure source is no less than the 
error In a calibrated pressure source mounted closer to the aircraft. 

CONCLUSIONS CONCERNING AIRSPEED CALIBRATION 

1. Use of the trailing anemometer device during a continuous calibration 
maneuver (gradual deceleration from maximum to minimum airspeed in 

a given configuration) produces the same results as data gathered 
during conventional steady-state (static) runs. The advantage of 
the continuous maneuver is a reduction by a factor of about 10 
in the flight time required to do an airspeed calibration. 

2. Airspeed calibration data from the tower-flyby method agree with 
data from the trailing anemometer methods. 

3. The effect of landing gear position on airspeed calibration is 
negligible for ATLIT. 

4. The effect of power-on airspeed calibration Is significant for ATLIT. 
The effect of power-on position error is greater with flaps down 
than with flaps up. 

5. The value in the use of a "long" instrumentation boom is questionable. 
Conventional practice dictates the use of standard lengths of nose 

or wing booms for pressure and airflow direction measurements. 
Typically, the boom length is prescribed as 1.5 body diameters 
(in the case of a nose boom) or 1.0 chord lengths (in the case of 



240 



a wing boom) in order to minimize position errors in pressure 
and airflow direction measurements. It can be argued that, if an 
installation is to be calibrated, the prescribed boom lengths can 
be shortened considerably. It can be shown that the calibration 
methods used have greater accuracy than the accuracy achieved in 
uncalibrated measurements from sensors on booms of the above 
lengths. For installations which are to remain uncalibrated, 
the above boom lengths would be appropriate, yielding pressure- and 
flow-direction measurements with minimal position errors. 

AN GLE OF ATTACK POSITION- ERROR CALIBRATION 

Calibration flights were performed to determine angle-of-attack 
position errors due to the airplane influence field. The test 
methods, data reduction, and results are discussed below. 

No additional instrumentation was required other than that 
described in Chapter 4.3. 

The calibration method used consists of equating indicated 
angle of attack (a') to pitch attitude angle (0) in straight and 
level flight conditions (a = 0). True angle^of attack (a) is defined 
for the present tests as the angle between the airplane longitudinal 
axis and the freestream velocity, or a = 0. 

An effective means of smoothing the scatter in the flight 
measured a and is achieved by plotting C. against both a and a' 
Since the shape of^ these curves can be cortfidently faired through the 



241 



242 
flight data (see figures B.8 and B.9), the angle-of -attack 
calibration curves are readily determined by plotting a vs. a' 
at constant C. 's (see figure B.IO). 

Calibrations were made for flap settings of 0°, lO*', and 30°. 
Tests were run with airplane-weight variations from 17300 N (3900 lb) 
to 18700 N (4200 lb). Center-of-gravity (CG) locations during the 
tests ranged from 12-percent mac to 15-percent mac. The varying CG 
locations result in different distributions of lift between the main 
wing and the horizontal tail; therefore, a given value of Cj^ could 

be generated by different angles of attack. The resulting effect 
on the flow field at the angle-of -attack vane is small and has been 
neglected. 

Results show a linear calibration correction between a' and a 
in the linear range of the C, vs a curves. The slope of the flaps-up 
calibration curve (figure B.IO) is 0.867. The figure also shows the 
effect of flap deflection on the angle-of-attack position-error 
calibration. 

The accuracy to which angle of attack may be calibrated by this 
method is limited to the accuracy of the vertical gyro. The gyro 
accuracy in the present tests is +0.7°. 



242 



243 
APPENDIX B - REFERENCES 

B.l Fisher, Bruce D.; Stough, H. Paul; and Kershner, David D.: 

Trailing Anemometer for Low-Airspeed Calibration. SAE Paper 

No. 760461, presented at the National Business Aircraft Meeting, 

Wichita, April 1976. 
B.2 Glauert, H.: Heavy Flexible Cable for Towing a Heavy Body 

Below an Aeroplane. Aeronautical Research Committee Reports 

and Memoranda No. 1592, London, 1934. 
B.3 Gracey, William: Measurement of Static Pressure on Aircraft. 

NACA TR 1364 (supercedes NACA TN-4184), 1957. 
B.4 Thompson, Floyd L.: The Measurement of Airspeed of Airplanes. 

NACA TN-616, 1937. 



243 



TABLE B.I.- SUMMARY OF STATIC PRESSURE POSITION-ERROR 
CALIBRATION METHOD ACCURACIES 



Method 


Source of Error 


Sample errors 
A V knots 


050 knots 


@140 knots 


Trailing 
Anemometer 


Anemometer Accuracy (Data System 
Noise and Wind-Tunnel Calibration) 


±1.1 


±1.1 




Induced Velocity (Fig. B.2) 


±0.1 


±0.1 




Flight-Measured Static Pressure 
(±43.2 psfa) 


±0.5 


±1.4 




Flight-Measured Dynamic Pressure 
(±1.0 psfd) 


±2.9 


±0.9 




Flight-Measured Temperature 








(±1.0O F) 

Root Mean Square Accuracy 
(one standard deviation) 


=^0 


=^0 


±1.4 kts 


±0.9 kts 


Tower 




@90 knots 


0170 knots 






Fly by 


Radar Angular Precision of ±0.2m 
(Fig. B.4) 


±0.4 


=^0 




Tower-Measured Static Pressure 
(±0.35 psfa) 


±0.6 


±0.3 




Tower-Measured Temperature 


=.0 


-0 




Flight-Measured Static Pressure 
(±3.5 psfa) 


±5.6 


±3.0 




Flight-Measured Dynamic Pressure 
(±1.0 psfd) 


±1.6 


±0.9 




Flight-Measured Temperature 








(±1° F) 


^0 


-0 


Root Mean Square Accuracy 
(one standard deviation) 


±2.6 kts 


±1.3 kts 



244 



FLAPS 


GEAR 
Up 


POWER 
Bleed Off 


V , AIRSPEED 
130 kts to stall 




METHOD 




0° 


Traill 


ng Anemometer - 


continuous run 


0° 


Down 


Bleed Off 


130 kts to stall 


Tra i 1 1 


ng Anemometer - 


continuous run 


0° 


Up 


Off 


130 kts to stall 


Traill 


ng Anemometer - 


continuous run 


0° 


Up 


For Level Flight 


75 kts 


Traill 


ng Anemometer - 


static run 


0° 


Up 


For Level Flight 


90 kts 


Traill 


ng Anemometer - 


static run 


0° 


Up 


For Level Flight 


110 kts 


Traill 


ng Anemometer - 


static run 


10° 


Up 


Bleed Off 


110 kts to stall 


Traill 


ng Anemometer - 


continuous run 


20° 


Up 


Bleed Off 


no kts to stall 


Traill 


ng Anemometer - 


continuous run 


30° 


Up 


Bleed Off 


110 kts to stall 


TraiV 


ng Anemometer - 


continuous run 


30° 


Up 


Approach 


110 kts to stall 


Trail - 


ng Anemometer - 


continuous run 


40° 


Up 


Bleed Off 


110 kts to stall 


Trailing Anemometer - 


continuous run 


0° 


Up 


For Level Flight 


87 kts 


Tower flyby 




0° 


Up 


For Level Flight 


105 kts 


Tower flyby 




0° 


Up 


For Level Flight 


130 kts 


Tower 


flyby 





TABLE B.2- CONFIGURATION/AIRSPEED COMBINATIONS FOR ATLIT 
STATIC PRESSURE-SYSTEM CALIBRATION TESTS. 



245 



IS3 




Figure B.I.- ATLIT trailing anemometer installation detail 



CO 

-3 



k = 



0^=2; A=6 




knots 



OTHER Cj_ AND A 



chart 



x3'Cjl 



.2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 



k =x/b 

X 



Figure B.2.- Induced velocity contours (from reference B.4) and 
trailing anemometer locations. 



CO 

*^ 

00 



METEOROLOGICAL 
TOWER 



"T" 

76.2 m 



61.0 m 



T" 

45.7 m 



X 




AN/FPQ-6 
RADAR 
NSTALLATION 



z (BY RADAR) 



3287.6 m 



•yyyyyy^yy^yyAyyyjyyyyyyy yyyyyyyy^^ 




Figure B.3.- Wallops tower flyby airspeed calibration dimensions. 



to 

CO 



.14 
.12 

.10 

CALIBRATION 
ERROR DUE 08 
TO ALTITUDE ' 

ERROR ^ .06 



.04 - 



.02 




40 



60 80 100 120 140 160 180 200 220 
MEASURED AIRSPEED V knots 



L L 

50 



1 



J 



100 



150 



200 



250 



MEASURED AIRSPEED V mph 



Figure B.4.- Effect of altitude error on tower flyby airspeed calibration. 



250 



O Trailing anemometer, static run 

D Tower fly by 
— Trailing anemometer, continuous run 




Indicated airspeed V ', knots 



Comparison of Tower Flyby and Trailing Anemometer{Static and Continuous Runs) 
Static Pressure Position Error Calibrations (6^=0 ) 

FIGURE B.5 



^ r>o 



0.1 



AP 



-0.1 



Power on, gear up, % = 

Power off, gear up, '^f = 0° 

Power on, gear up, °f = 30 



Qy. 



it. Power off, gear up, '^f = 30 [jt-ttt 




nm 



Indicated airspeed V ', knots 

Effect of Power on Static Pressure Position Error Calibrations 
(Trailing Anemometer Continuous Runs) 

FIGURE B.6 



250 




Indicated airspeed V^,', knots 

Effect of Flap Deflection on Static Pressure Position Error Calibration 
(Trailing Anemometer Continuous Runs) Power on 

FIGURE B.7 



251 




a, deg 

Figure B. 8.- Flight- test lift data for true 
(geometric) angles of attack 
(6^ = o'^ (spoiler leak path 
sealed), 10°, 30°). CG =^ 15% mac 



252 




Figure B.9.- Flight- test lift data for indicated angles-of-attack. 

(6^ = 0° (spoiler leak path sealed) 10°, 30°) CG « 15%mac 



253 



20 



d' , deg 



15 



10 



-5 




-5 



10 



15 



20 



a, deg 

(a) Flaps up 
Figure B.IO- Angle-of-attack position-error calibrations 



254 



20 



15 



10 



a' » deg 



-5 




-5 



5 10 
a, deg 
(b) Flaps 10° 
Figure B,10- continued 



15 



255 



20 



15 



10 



a' , deg 



-5 



-10 




-10 



-5 



10 



15 



a, deg 
(c) Flaps 30° 

Figure B.IO- Concluded. 



256 



APPENDIX C 



PREDICTION OF ROLL DAMPING DERIVATIVES 



257 



APPENDIX C 

PREDICTION OF ROLL DAMPING DERIVATIVES 

The purpose of this appendix is to present the method,, data, and 
results for analytical predictions of ATLIT roll damping derivatives, 
C, ) for several combinations of lift coefficient and flap deflection. 

The bulk of the computational work for this appendix was performed by 
Mr. Bradley J. Vincent and his contributions are gratefully acknowledged. 

Method of Analysis .- The method used to predict roll damping 
derivatives is from reference C.l. The method is incorporated in a 
computer program described in reference C.2. This method does not provide 
for airplane configurations with wing-mounted engine nacelles or with 
flaps deflected. The methods by which these cases were handled are 
described below. 

Effect of Engine Nacelles .- A sample calculation of the contribution 
of the engine nacelle to the total airplane roll damping will show the 
effect to be smal 1 . 

The rolling moment coefficient of the nacelle alone is computed 
for the conditions and assumptions presented in figure C.l. 

Computing the increment of rolling moment coefficient contributed 
by the nacelle yields: 



^^1 n 


= L, .d ^ 


«» " Sy 


d 




% h ^ 


b 


1 . n 


= -0.00018 







258 



Nondimensionalizing, with. respect to airplane, pb/2V yields the 
nacelle contribution to the airplane roll damping derivative: 



AC, „ = ^t^l 
1 . n 



^ 2V ' 

AC, = -0.0034 
p, n 

This value amounts to less than 1 percent of the estimated total airplane 

roll damping at low-lift coefficients (flaps up) and is neglected in the 

final analysis of C^ . 

P 

Effect of Flap Deflections.- To estimate C, with flaps down, 

P 

the geometry of the wing was recomputed at each flap deflection (see 
table C.l). Thus, the assumption is made that roll damping with flaps 
down may be estimated by considering the flap deflection simply as a 
change in the wing area, aspect ratio, and taper ratio. 

The sources for additional inputs to the program are explained 
as follows: 

ot The computer program requires true (geometric) angles 
of attack. First, C. is computed for the condition 

of interest. Then, a is obtained from the flight- test 
C, ys a curves of Appendix B and Chapter 5. 



259 



C, The program requires two-dimensional lift curve slopes. 
a 

These are based on wind-tunnel test data (reference C.3) 

for the GA(W)-1 airfoil with the Fowler flap. These 

values for C^ are summarized in figure C.2. 

a 

Cpj The airplane zero- lift drag coefficient was estimated 



based on preliminary flight- test results. Based on 
wind-tunnel reflection plane-test data for the ATLIT 
wing (reference C.4), increments were added to account for 
increases in zero lift drag at increased flap deflections 
The estimated values for C^ follow: 





'f 






\ 


0° 


0.040 


10° 






0.050 


20° 






0.083 


30° 






0.136 


40° 






0.181 


The final estimates of 


airpl 


ane 


roll damping derivatives appear in 



figure C.2. The estimates are presented for varying lift coefficients 
with flap settings of 6^ = 0^, 10°, and 30 . The trends shown in the 
figure for decreasing roll damping with increasing C. are expected 

due to decreasing lift curve slopes at higher angles of attack 

(higher C. ). The large increase in flaps-down roll damping is expected 

due to both increases in C, and changes in the wing planform with the 

a 

flaps deflected. 



260 



APPENDIX C 

REFERENCES 

C.l Hoak, D. E.; Ellison, D. E., et al : USAF Stability and Control 
Datcom. Flight Control Division, Air Force Flight Dynamics 
Laboratory, Wright Patterson Air Force Base, Ohio, 45433, 1972. 

C.2 Smetana, F. 0.; Summey, D. C; and Johnson, W. D.: Flight Testing 
Techniques for the Evaluation of Light Aircraft Stability 
Derivatives, a Review and Analysis. NASA CR 2016, 1972. 

C.3 Wentz, W. H., Jr.; and Seetharam, H. C: Development of a 

Fowler Flap System for a High-Performance General Aviation 
Airfoil. NASA CR 2443, 1974. 

C.4 Wentz, W. H., Jr.; and Volk, C. G., Jr.: Reflection-Plane Tests 
of Spoilers on an Advanced Technology Wing with a Large 
Fowler Flap. NASA CR 2696, 1976. 



261 



TABLE C.I.- WING GEOMETRY WITH FLAPS DEFLECTED 



Flap 
Deflection 


S, m ^ (ft^) 


A 


X 


6^= 0° 


14.4 (155.0) 


10.32 


0.5 


6f = 10° 


16.7 (179.6) 


8.91 


0.5 


6f = 20° 


17.1 (183.9) 


8.70 


0.5 


6f = 30° 


17.3 (186.1) 


8.60 


0.5 


6f = 37° 


17.0 (183-2) 


8.73 


0.5 



262 




Center of nacelle lift 



V = 174 knots (200 mph) 

p = 1.0 rad/sec 

d = 1.85m (6.08 ft) 
b ^ 12.2m (40 ft) 



■a.n 



= 0.86 rad 



-1 



d.p 
n V 

S^ = 1.23m^ (13.27 ft^) 

S^ = 14. 4m^ (155 ft^) 

W 



Figure C.I.- Conditions and assumptions for estimating engine nacelle 
contribution to airplane roll damping derivative. 



263 



ts3 

4^- 



0.15 



0) 



0) 

■a 



0.10 



0.05 




a, deg 



Figure C.2.- Variation of two-dimensional lift curve slope with angle of attack 
(6f = 0°, 10°, 20°, 30°, 40°). 



0.5 



-C 




1.0 2.0 

Airplane lift coefficient, Cl 



3.0 



Figure C.3.- Predicted airplane roll damping derivatives 

(6f = 0°, lOO, 30°) 



NASA-Langley. 1977 CR-2832 



265